Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Medicine 1342

Microarray Technology forGenotyping in Pharmacogenetics

BY

ULRIKA LILJEDAHL

ACTA UNIVERSITATIS UPSALIENSISUPPSALA 2004

To My Family

iv

Supervisor: Ann-Christine Syvänen, Professor Molecular Medicine Department of Medical Sciences Uppsala University Sweden

Faculty opponent: Denis Grant, Professor and Chair Department of Pharmacology Institute for Drug Research University of Toronto Canada

Review board: Magnus Ingelman-Sundberg, Professor Molecular Toxicology The Institute of Environmental Medicine Karolinska institutet Sweden

Elena Jazin, Docent Behavioral Genetics Department of Evolutionary Biology Uppsala University Sweden

Maria Lagerström-Fermér, Docent Department of Molecular Sciences AstraZeneca R&D Mölndal Sweden

v

LIST OF PUBLICATIONS

This thesis is based on the following publications, which will be referred to in the text by their roman numerals:

I Katarina Lindroos, Ulrika Liljedahl, Mirja Raitio and Ann-Christine Syvänen. Minisequencing on oligonucleotide microarrays: comparison of immobilisa-tion chemistries. Nucleic Acids Research, 2001, 29(13): e69.

II Ulrika Liljedahl, Julia Karlsson, Håkan Melhus, Lisa Kurland, Marie Linders-son, Thomas Kahan, Fredrik Nyström, Lars Lind and Ann-Christine Syvänen. A microarray minisequencing system for pharmacogenetic profiling of anti-hypertensive drug response. Pharmacogenetics, 2003, 13(1): 7-17.

III a) Lisa Kurland, Ulrika Liljedahl, Julia Karlsson, Thomas Kahan, Karin Malmqvist, Håkan Melhus, Ann-Christine Syvänen and Lars Lind. Angiotens-inogen gene polymorphisms: Relationship to blood pressure response to anti-hypertensive treatment. American Journal of Hypertension, 2004, 17(1):8-13.

b) Ulrika Liljedahl, Lisa Kurland, Lars Berglund, Thomas Kahan, Ann-Christine Syvänen and Lars Lind. Single nucleotide polymorphisms in the apolipoprotein B and low density lipoprotein receptor genes affect response to antihypertensive treatment. Submitted.

c) Ulrika Liljedahl, Thomas Kahan, Karin Malmqvist, Håkan Melhus, Ann-Christine Syvänen, Lars Lind and Lisa Kurland. Single nucleotide polymor-phisms predict the change in left ventricular mass in response to antihyper-tensive treatment. Submitted.

IV Mona Fredriksson, Gisela Barbany, Ulrika Liljedahl, Monika Hermanson, Matti Kataja and Ann-Christine Syvänen. Assessing hematopoietic chimerism after allogeneic stem cell transplantation by multiplexed SNP genotyping us-ing microarrays and quantitative analysis of SNP alleles. Leukemia, 2004, 18(2):255-266.

V Ulrika Liljedahl, Mona Fredriksson, Andreas Dahlgren and Ann-Christine Syvänen. Detecting imbalanced expression of SNP alleles by minisequencing on microarrays. Submitted.

The original publications are reprinted by permission of Oxford University Press, Lip-pincott Williams&Wilkins, Elsevier Publishing and Nature Publishing group.

vi

ABBREVIATIONS

ASO allele specific oligonucleotide BP blood pressure cDNA complementary DNA DBP diastolic blood pressure dNTP deoxynucleoside triphosphate ddNTP dideoxynucleoside triphosphate gDNA genomic DNA HapMap the Haplotype Mapping Project LD linkage disequilibrium LNA locked nucleic acid LVH left ventricular hypertrophy LVM left ventricular mass OLA oligonucleotide ligation assay PCR polymerase chain reaction PNA peptide nucleic acid RAAS renin-angiotensin-aldosterone systemSBP systolic blood pressure SNP single nucleotide polymorphism Tm melting temperature

vii

CONTENTS

INTRODUCTION ....................................................................................................9

HUMAN GENOME RESEARCH ...................................................................................9Sequence variation ..........................................................................................10

Single nucleotide polymorphisms............................................................................10Haplotypes and linkage disequilibrium...........................................................11

The Haplotype Mapping Project..............................................................................11SNPs in disease genetics .................................................................................12

Monogenic disease ..................................................................................................12Multifactorial disease ..............................................................................................12Identifying markers underlying genetic disease.......................................................13

PHARMACOGENETICS............................................................................................14Pharmacokinetics ............................................................................................15Pharmacodynamics .........................................................................................16Pharmacogenetics of antihypertensive treatment............................................16

Hypertension ...........................................................................................................16Left ventricular hypertrophy....................................................................................17

Genotyping in pharmacogenetics ....................................................................18ARRAY TECHNOLOGY ...........................................................................................19

Array based genotyping...................................................................................19DNA polymerase assisted assays.............................................................................19

Single nucleotide primer extension ....................................................................19Allele specific primer extension.........................................................................21

Ligation based assays ..............................................................................................22Hybridisation based assays ......................................................................................23Towards PCR free genotyping.................................................................................24

Detection .........................................................................................................25Microarray preparation ..................................................................................27

Array surfaces..........................................................................................................27In situ synthesis of oligonucleotides ........................................................................27Immobilisation of pre-synthesised oligonucleotides................................................28

Quantitative SNP analysis...............................................................................29

PRESENT STUDY .................................................................................................31

INTRODUCTION .....................................................................................................31AIMS.....................................................................................................................31MATERIALS AND METHODS ..................................................................................32

Study samples ..................................................................................................32The SILVHIA trial ..................................................................................................32

SNP selection and primer design ....................................................................33Polymerase chain reaction..............................................................................34Preparation of microarrays.............................................................................34

viii

Minisequencing in the microarray format.......................................................35Minisequencing in the microtiter plate format ................................................37Signal detection ...............................................................................................37Genotype assignment.......................................................................................38Statistical analysis...........................................................................................38

RESULTS AND DISCUSSION ....................................................................................38Microarray based genotyping .........................................................................38

Immobilisation of oligonucleotides .........................................................................38Genotyping with single-colour fluorescence detection............................................41Genotyping with four-colour fluorescence detection...............................................43Accuracy and sensitivity of allele quantification .....................................................43Detection of alleles in mixed DNA samples ............................................................46Detection of allelic imbalance .................................................................................47SNP selection ..........................................................................................................48

Applications.....................................................................................................49Pharmacogenetics of antihypertensive response......................................................49

CONCLUDING REMARKS .................................................................................54

ACKNOWLEDGEMENTS ...................................................................................55

REFERENCES........................................................................................................57

9

INTRODUCTION

Human genome research In 1944, Avery and colleagues identified the genetic material as deoxy-ribonucleic acid, DNA, when studying the transfer of genetic material in bacteria (Avery et al. 1944). Fifty years following Watson and Crick’s presentation of the helical structure of DNA (Watson and Crick 1953), the reference sequence of the three billion nucleotides constituting the human genome was announced in April 2003 with 99% coverage of the gene-containing regions at an accuracy of 99.99%. This was made possible through the work of the public Human Ge-nome Project (Lander et al. 2001) and the private Celera genome pro-ject (Venter et al. 2001).

Methods for the sequencing of DNA had been presented in 1977 by two independent research groups. The dideoxynucleotide method, better known as “Sanger sequencing”, is the technique used for DNA sequencing today (Sanger et al. 1977), while the chemical degradation method by Maxam and Gilbert (Maxam and Gilbert 1977) is rarely used for sequencing purposes. However, this latter technique forms the basis for mutation detection by “chemical cleavage of mismatch” (Cotton et al. 1988).

Another breakthrough in the field of genetic research came in 1985 when the method for the amplification of specific regions of the ge-nome sequence by the polymerase chain reaction (PCR) was presented (Saiki et al. 1986; Mullis and Faloona 1987). PCR is today a funda-mental basis for many genotyping methods for amplification of the genomic regions of interest. The amplification provides sufficient sensi-tivity and specificity in detecting a genetic variant among the 3·109

base pairs of DNA that constitute the human genome. Protein coding regions constitute less than 2% of the genome and

the function of most of the other 98% remains to be elucidated (Levine and Tjian 2003). It has been proposed that the human genome contains approximately 30,000 protein coding genes (Katsanis et al.2001; Lander et al. 2001; Levine and Tjian 2003), which is somewhat surprising since this is only about twice as many as in less complex organisms such as the worm C. elegans or the Drosophila fly (Levine and Tjian 2003). It is therefore reason to believe that the difference in

10

complexity among organisms is due to alternative splicing at the level of transcription (Maniatis and Tasic 2002; Roberts and Smith 2002). About 30-60% of the human genes undergo alternative splicing result-ing in several protein coding mRNAs from each gene (Lander et al.2001; Modrek and Lee 2003; Sorek et al. 2004).

Information on sequence variation among individuals and popula-tions has also been provided by the human genome sequencing project. The extent of the variability in the genome and its effect on disease phenotype or variability in responsiveness to drug treatment is now being evaluated.

Sequence variation Single nucleotide polymorphisms The DNA sequence for multiple individuals is identical to 99.9%. The variable 0.1% is mainly single nucleotide polymorphisms (SNPs) which can explain a large fraction of the inter-individual differences in pheno-type between population and individuals (Sachidanandam et al. 2001; Marth et al. 2003).

Depending on the location of a SNP, the phenotypic consequences may differ. SNPs within the protein coding regions of a gene may alter an amino acid (i.e. non-synonymous change) and thus change the structure and function of the protein. The nucleotide change can also result in a synonymous change, having no effect on the amino acid composition of the protein. SNPs in regulatory regions of a gene are gaining more and more interest today since these may affect the bind-ing of transcription factors and thereby the expression level of a gene which can result in a disease phenotype (Hudson 2003). The majority of the SNPs are located in the non-coding parts of the genome where they have no direct effect on a phenotype. These are useful as genetic markers and may be linked to functionally important SNPs in the cod-ing or regulatory regions of the gene.

More than nine million SNPs are available in the public SNP data-base at NCBI (dbSNP, build 120, March 18th 2004; http://www.ncbi.nlm.nih.gov/SNP). About half of these have been confirmed by multiple reports to the database, observation of the al-leles in at least two chromosomes or by allele frequency and/or geno-typing data.

The usefulness of SNPs in genetic research is due to their frequent and even distribution in the genome, their genetic stability and usually biallelic feature making them suitable for high throughput automated genotyping assays. Much effort has been made trying to identify the genetic variants responsible for, or describing susceptibility to, human

11

disease; this will be discussed further in the section “SNPs in disease genetics”.

Haplotypes and linkage disequilibrium Haplotypes are composed of a series of alleles, at linked loci on a single chromosome, which are inherited together in a non-random pattern. This association between alleles is termed linkage disequilibrium (LD) and occurs for markers located close to each other on a chromosome. It has been proposed that haplotypes are inherited in block-like struc-tures of low diversity and high LD, separated by regions where LD is broken down due to recombination events (Daly et al. 2001; Goldstein 2001; Patil et al. 2001; Gabriel et al. 2002; Phillips et al. 2003). Such “recombination hotspots” have been identified in a few genomic re-gions, for example in the MHC class-II region and the -globin locus (Smith et al. 1998; Jeffreys et al. 2001). However it is not clear whether this pattern is a general feature of the genome. The extent of haplotype blocks and presence of recombination hotspots is currently highly debated (Cardon and Abecasis 2003; Phillips et al. 2003; Wall and Pritchard 2003). In addition to the differences in LD between ge-nomic regions, there is variability between populations, with a lower level of LD in Africans compared to the European and Asian popula-tions (Reich et al. 2001; Gabriel et al. 2002; Shifman et al. 2003; Stumpf and Goldstein 2003). The variability in LD can be explained by mutation rates, population growth and genetic drift, or admixture as reviewed in (Ardlie et al. 2002; Tishkoff and Verrelli 2003). Knowl-edge of the pattern of linkage disequilibrium throughout the human genome would be helpful in the identification of common genetic vari-ants involved in complex disease (Cardon and Abecasis 2003; Tishkoff and Verrelli 2003) (section “Identifying markers underlying genetic disease”).

The Haplotype Mapping Project As a joint effort by a consortium consisting of leading academic and commercial scientists in Japan, United Kingdom, Canada, China, Nige-ria and the United States, the work on constructing a haplotype map (HapMap) of the human genome was launched in October 2002 (HapMapConsortium 2003). The major goal of the HapMap project is to characterise the pattern of LD and underlying haplotypes in our genome. Samples from Nigerian Yoruba, Han Chinese, Japanese and a US Utah population of Northern and Western European descent will be analysed revealing the ethnic variability in LD patterns and haplo-type blocks (HapMapConsortium 2003; Deloukas and Bentley 2004).

12

All information generated will be made publicly available and thus this map will provide guidelines for future linkage-based and genome-wide association studies to elucidate the sequence variants predisposing to common multifactorial diseases or variable drug responses, as well as providing useful information for the development of diagnostic tools.

SNPs in disease genetics

Monogenic disease Diseases following a Mendelian recessive or dominant pattern of in-heritance, more commonly called monogenic diseases, are caused by genetic variation in a single gene most often leading to an absent or modified protein (Peltonen and McKusick 2001). The monogenic dis-eases are rare and identification of their genetic cause has been facili-tated by the availability of genome wide genetic maps of microsatellite markers. Information on Mendelian disorders is catalogued in the OMIM database (Online Mendelian Inheritance in Man, http://www.ncbi.nlm.nih.gov/Omim) where 1500 phenotypes linked to disease are described (March 12, 2004). A few examples on mono-genetic diseases and their causative genes are given in Table 1.

Multifactorial disease Considering non-Mendelian traits, the disease phenotype is the result of genetic variation in one or several susceptibility genes in a complex interplay with environmental factors such as age, weight or smoking habits. These complex, multifactorial disorders are more common than the monogenic disorders, but determination of the extent of genetic contribution to disease and the identification of the actual disease genes has been complicated (Peltonen and McKusick 2001). There are however several examples of complex diseases for which susceptibility loci have been identified and some of them are listed in Table 1.

The localisation and identification of a disease gene can be per-formed by linkage and association studies. For Mendelian traits, linkage analysis of genetic markers co-segregating with a disease in affected families has been successful. Linkage detects inheritance of alleles at loci close together on the same chromosome because recombination rarely occurs between these loci during meiosis (Botstein and Risch 2003). This approach has provided limited success in finding the genes related to common, complex disorders with a complicated mode of inheritance. Instead association or LD studies have been proposed (Cardon and Abecasis 2003). Association studies involve the genotyp-ing of markers in a genomic region, or a candidate gene, and comparing the allele frequencies of individuals with the disease phenotype to that

13

of controls. However, the results are uncertain due to the fact that only a fraction of the markers in a region has been used, and thus only part of the genetic variation has been analysed.

Table 1. Examples of monogenic and multifactorial diseases and associated genetic loci.

Disease Genetic component References

Monogenic diseases

Cystic fibrosis Deletion of amino acid 508 most common in CF patients

(Kerem et al. 1989)

Huntington disease Expansion of CAG repeat in the huntingtin gene

(Huntington'sDiseaseCol-laborativeGroup 1993)

Multifactorial dis-eases

Asthma ADAM33 risk factor (Van Eerdewegh et al.2002)

Late onset Alz-heimer’s disease

ApoE 4 risk factor (Corder et al. 1993)

Crohn’s disease Genetic variants and haplotypes in CARD15 (NOD2)

(Hugot et al. 2001; Oguraet al. 2001; Croucher et al. 2003)

Crohn’s disease and ulcerative cholitis

Haplotype in the IBD5 gene (Rioux et al. 2001; Gial-lourakis et al. 2003)

Myocardial infarc-tion

Increased susceptibility by functional SNPs in lymphotoxin A gene

(Ozaki et al. 2002)

Myocardial infarc-tion and stroke

5-lipoxygenase activating protein (Helgadottir et al. 2004)

Type 2 diabetes Calpain-10 association in Mexican Americans and Northern Europeans

(Horikawa et al. 2000)

Identifying markers underlying genetic disease Linkage disequilibrium mapping relies on the identification of allelic association between a genetic marker and the disease loci involving both linkage and association (Clark 2003). A major goal for the Hap-Map project is to characterise the pattern of LD and identify haplotype tagging SNPs that define the haplotype blocks (Johnson et al. 2001; HapMapConsortium 2003). These common htSNPs would be useful for identifying disease associated markers according to the common disease/common variant hypothesis. This hypothesis predicts that the genetic risks for common diseases will be associated with disease alleles of relatively high frequencies (Reich and Lander 2001; Pritchard and

14

Cox 2002). Instead, if multiple rare alleles would predispose to com-plex diseases, the use of htSNPs for LD mapping would prevent the detection of the disease associated genetic variants (Tishkoff and Ver-relli 2003). Our understanding of the pattern of LD is still limited and the influence of the block-like structure and marker density in LD mapping for whole genome association studies remains to be eluci-dated (Cardon and Abecasis 2003; Ke et al. 2004).

Pharmacogenetics The concept of pharmacogenetics dates back fifty years to observations of unexpected drug responses for which a heritable component was suggested due to a familial pattern of the drug response (Hughes et al.1954; Alving et al. 1956; Lehmann and Ryan 1956). It was not until later that we were able to identify the genetic variations causing these findings. The study on the effect of genetic components on drug re-sponse (pharmacogenetics) is divided into pharmacokinetics and pharmacodynamics. Pharmacokinetics describes genetic factors of the absorption, distribution, metabolism and excretion of a drug, while pharmacodynamics is defined as the genetic factors describing the pharmacological effect of a drug in the body and subsequently the ef-fect on the phenotype (Figure 1).

The use of pharmacogenetics in a clinical setting holds the promise of individualisation of drug treatment, reducing the risk of side effects and time for treatment and thus a more cost effective healthcare (Evans 2003; Johnson 2003; Weinshilboum 2003). A public repository of genotype and phenotype information relevant to pharmacogenetics is available in the Pharmacogenetics and Pharmacogenomics Knowl-edge Base (PharmGKB, http://www.pharmgkb.org). Researchers are encouraged to deposit their pharmacogenetic findings in the database to make the information available in a structured way and promote sharing of information to facilitate research in the field (Klein and Altman 2004).

15

Figure 1. The processes that take place when a drug enters the body is de-scribed by pharmacokinetics. The absorption, distribution, metabolism and elimination of the drug determine the level of the drug in the blood. Genetic variants in genes encoding the metabolising enzymes can result in a toxic ef-fect of the drug because of an impaired metabolism. On the other hand, if the metabolism is too extensive the dose commonly used might lack effect. Phar-macodynamics describes the effect of the drug in the body and involves the binding of the drug or a metabolite to a receptor that mediate its effect. Ge-netic variation in the genes encoding the receptors may lead to an altered expression, structure or function of the receptor and thus reducing the effi-cacy of the drug. (Figure adapted from: (Turner et al. 2001).)

Pharmacokinetics Among the most well characterised drug metabolising enzymes is the cytochrome P450 system (CYP450), responsible for metabolism of numerous drugs in clinical use today. This oxidative enzyme system is thought to have evolved in rodents as a means to enable them to eat and metabolise compounds in plants that would otherwise be toxic. In humans, the majority of the CYP enzymes are primarily expressed in the liver. The CYP450 enzymes have been identified to harbour ge-netic variants important in the pharmacogenetics of drug response. CYP2D6 is the isoform most extensively studied, responsible for the metabolism of numerous neuroleptics and antidepressant drugs. An-other well characterised isoform is the CYP2C9 for which individuals with a slow metaboliser phenotype suffers from increased risk of bleed-ing when treated with the anticoagulant warfarin (Aithal et al. 1999; Wadelius et al. 2004). Information on the genetic variability of differ-ent CYP450 isoforms is available at www.imm.ki.se/CYPalleles and recently reviewed in (Ingelman-Sundberg 2004).

Other pharmacokinetically important enzymes involved in the drug metabolism have been identified. Two of the most extensively studied are the N-acetyltransferase (NAT) and the cytosolic thiopurine me-

16

thyltransferase (TPMT). N-acetyltransferases (NAT) is responsible for the metabolism of several antiarrhythmics and antidepressants. Com-mon genetic variants have been identified to affect the plasma concen-trations of such drugs as reviewed in (Grant et al. 2000). TPMT is in-volved in the metabolism of thiopurines used in cancer treatment, and the enzymatic activity is highly polymorphic which affects the treat-ment efficiency and risk of toxic reactions (Evans and Relling 1999; Seidman 2003).

PharmacodynamicsPharmacogenetic research is now turning towards the field of pharma-codynamics and the study of drug targets, to evaluate the effect of ge-netic variation on the variability in drug response among individuals. In this relatively new field of pharmacogenetic research, the studies so far have been small and without any conclusive results. The best studied areas of pharmacodynamics are cardiovascular disease and asthma as reviewed by Johnson etal. (Johnson and Lima 2003).

An example of pharmacodynamically important genetic variation is found for asthma. A genetic variation leading to the change from gly-cine to arginine at aminoacid 17 in the 2-adrenergic receptor has been associated with the response to beta-agonist therapy (Liggett 2000). Genetic variants in genes encoding components of the renin-angiotensin-aldosterone system (RAAS) have been associated to the blood pressure change during treatment of hypertension (section “Hy-pertension” and Studies II and IIIa-b).

Pharmacogenetics of antihypertensive treatment HypertensionEssential hypertension has a multifactorial aetiology that includes the interaction of many genetic and environmental factors acting through the intermediate systems regulating blood pressure control (Guyton et al. 1981; Lifton et al. 2001). Hypertension has a prevalence of 25-35% in the adult population in developed countries (Staessen et al. 2003). The heritability of blood pressure has been studied in twins and family material and been determined to 30-70% (Pausova et al. 2001; Fagardet al. 2003).

About half of the individuals with hypertension are treated and only 50% of those have adequate blood pressure control although several drugs are used (WHO 2003; Mein et al. 2004). This problem can be explained by the heterogeneity of hypertension as a disease entity as well as the individuality of the drug response. It is not possible today to

17

determine in advance how the individual patient would respond to treatment by a specific drug. Metabolic parameters such as renin levels have been suggested as predictors for drug response but none has been identified so far. Identifying genetic variants that could be used as pre-dictors for the outcome of antihypertensive treatment would help identifying the best drug for the individual patient, thus reducing the cost for the treatment and the care time (Turner et al. 2001; Turner and Boerwinkle 2003). With a better treatment of hypertension with more patients treated to goal, the incidence of myocardial infarction and stroke will be reduced. A better treatment will lead to a reduction in morbidity and mortality, as well as a reduction in the cost for treat-ment.

The search for important sequence variants to be used in the phar-macogenetics of antihypertensive response has mainly focused on can-didate genes of the RAAS. Many of the studies have been limited by small number of SNPs and samples analysed, as reviewed by Koop-mans etal. (Koopmans et al. 2003). Ethnic variation in drug response as well as different phenotypes that in some cases are not adequately de-fined may also contribute to the inconclusive results. The amino acid change from arginine to glycine at amino acid residue 389 in the 1-adrenergic receptor has been associated to the change in systolic blood pressure during atenolol treatment (Sofowora et al. 2003). Genetic variants in the angiotensinogen gene (StudyIIIa; (Kurland et al. 2004), angiotensin converting enzyme (Kurland et al. 2001) and the aldoster-one synthase (Kurland et al. 2002a) genes have been associated to the response to antihypertensive treatment with the angiotensin II type 1 receptor antagonist irbesartan or the 1-adrenergic receptor blocker atenolol. This has been evaluated in Studies II and IIIa-b.

Left ventricular hypertrophy Left ventricular hypertrophy (LVH) is one of the most powerful, in-dependent risk factors for cardiovascular morbidity and mortality in the general population (Levy et al. 1990), and in hypertensive indi-viduals in particular (Koren et al. 1991; Schillaci et al. 2000). LVH is a multifactorial disease with hemodynamic, neurohormonal and genetic factors as the cause. LVH is strongly determined by inherited factors, and has been proposed to contribute to the development of hyperten-sion (Devereux et al. 1991; Koren et al. 2002).

There are indications that angiotensin II type 1 receptor antagonists may have advantages over other antihypertensive drugs in reducing LVH in hypertensive patients (Malmqvist et al. 2001; Dahlof et al.2002). The effect of genetic variation in candidate genes of blood pres-sure regulation on the reduction in LVH upon antihypertensive treat-ment has been evaluated in Study IIIc.

18

Genotyping in pharmacogenetics Adverse reactions to drugs are the fourth to sixth leading cause of death in the United States. Of all individuals admitted to hospital 6.7% are suffering from adverse reactions to drug treatment (Lazarou et al.1998). In addition to a more efficient treatment, pharmacogenetics has the potential to reduce the number of adverse drug reactions and guide in the administration of drugs and thus provide a more cost efficient health care. There are two possible methods to perform pharmacoge-netic testing, either through phenotyping or genotyping. The former involves administering a drug that is metabolised by the pathway of interest and analyse the drug and its metabolites to assess the enzyme activity of an individual, thus a pharmacokinetic test. As development of technologies for genotyping and identification of clinically important genetic variants in drug metabolising enzymes and drug receptors pro-ceeds, the determination of an individual’s phenotype is facilitated. By analysis of the genotype of the variants of interest, the phenotype can be predicted from a blood sample.

Although numerous genetic variants in pharmacokinetically and pharmacodynamically important genes have been identified, only a few are being used in practice in clinical medicine today. In Sweden, geno-typing is recommended for two drugs prior to administration, these are the antiarrhythmic agent propafenon, metabolised by CYP2D6; and the tuberculostatic agent isoniazid, metabolised by NAT2. Table 2 presents some examples of the application of pharmacogenetic analysis in different research areas.

Table 2. Examples of pharmacogenetic studies.

Application Genes analysed References

Cancer susceptibility and phar-macogenetics

42 genes and 93 SNPs

(Landi et al. 2003)

6-mercaptopurine and azathio-prine in inflammatory bowel disease

Thiopurine methyl-transferase

(Seidman 2003)

Treatment of acute lymphoblas-tic leukaemia in children

Thiopurine methyl-transferase

(McLeod et al. 2000)

Pharmacogenetic testing in the clinic to identify poor and ultra-rapid metabolisers

Primarily CYP2D6 AmpliChip CYP450, Roche Diagnostics

Outcome of antidepressant treat-ment

Serotonin locus and CYP2D6

(Murphy et al. 2003)

19

Array technology With the development of array based technologies, research has shifted from the analysis of individual genes to highly parallel assays. The con-cept of oligonucleotide microarrays, a set of oligonucleotides or DNA molecules immobilised on a solid support, was first proposed to be used in sequencing by hybridisation either with immobilised target DNA (Drmanac et al. 1989; Drmanac et al. 1993) or immobilised oli-gonucleotide probes (Southern et al. 1992). The microarray format has been extensively utilised in expression profiling of a large number of mRNA species for a relatively small number of samples (Schena et al.1995; Arava et al. 2003). To understand the phenotypic effect of ge-netic variation, such as SNPs, many samples and multiple markers need to be analysed. This requirement is driving microarray technology further. The next step would be the multiplexed analysis of protein by microarray technology similar to that used in RNA and DNA research, for example using immobilised antibodies or specific protein ligands (Eickhoff et al. 2002)

Because of the miniaturised format of array-based technologies, the cost for reagents is reduced as well as the amount of work needed per gene or gene product. However, an enormous amount of data is gener-ated from microarray experiment, creating new demands on handling of the data in databases and for statistical analysis.

In the following sections, methodology for genotyping genetic vari-ants in various array based formats will be discussed.

Array based genotyping Three major reaction principles form the basis of the genotyping tech-nologies available today; these are primer extension, ligation and hy-bridisation. Many of the methods for SNP genotyping rely on PCR amplification of the sequence of interest prior to allele determination. However, difficulty in performing highly multiplex PCR reactions needed for the parallel analysis of many SNPs is a major bottle neck in most methods used today. Methods with the amplification step subse-quent to the allele recognition step, to increase the signal for detection or methods without PCR, are gaining more and more interest.

DNA polymerase assisted assays Single nucleotide primer extension The DNA polymerase assisted single nucleotide primer extension reac-tion is widely used and known by several synonyms such as minise-quencing (Syvanen et al. 1990), nested genetic bit analysis (N-GBA) (Head et al. 1997), single base extension (SBE) (Fan et al. 2000), single

20

nucleotide extension (SNE) (Fortina et al. 2000) and arrayed primer extension (APEX) (Kurg et al. 2000). The simplicity of the reaction and the need of only one detection primer to identify the allelic vari-ant, makes the reaction easy to establish and suitable for multiplexing to a high level. Microarray based minisequencing has shown a nine-fold better genotype discrimination compared to hybridisation with allele specific oligonucleotide probes (Pastinen et al. 1997).

The primer extension reaction mimics the replication of genetic in-formation during DNA synthesis in the cells and relies on the high sequence specificity in allele discrimination by the DNA polymerase. In the minisequencing assay, an oligonucleotide designed to anneal only adjacent to the SNP position is extended with a dNTP or ddNTP by the action of the DNA polymerase (Figure 2A) (Syvanen et al.1990; Syvanen et al. 1993). Many varianys of genotyping by primer extension in a microarray based format are available, and a rough summary is presented in Table 3. There is a constant improvement in reaction conditions due to new enzymes and nucleotide analogues. The choice of reaction procedure, whether performing the extension in situon the slide or in solution, has an impact on the results. The solution phase reaction is thought to give higher signals due to a more favour-able kinetic environment of the components of the extension (Fortinaet al. 2000; Tsuchihashi and Dracopoli 2002). The solution phase sys-tem may however suffer from more dimer formations between detec-tion primers for different SNPs, compared to the assay with the detec-tion primers immobilised on the surface.

Figure 2. Reaction principles for DNA polymerase assisted SNP genotyping. The detection of the A-allele of an A to G transition is illustrated. In the minisequencing single nucleotide primer extension reaction (A), a primer designed to anneal immediately adjacent to the SNP position is extended with a nucleotide complementary to that of the SNP position. In the allele specific primer extension reaction (B), the 3’-nucleotide of the primer is complemen-tary to the SNP allele. A perfect match allows for the extension of the primer with a labelled nucleotide analogue.

21

Primer extension reactions have also performed well in solution with colour coded microbeads (Luminex) (Cai et al. 2000; Chen et al.2000). Although this is not a microarray system as such, it represents a system allowing parallel analysis of many polymorphisms simultane-ously due to spectral individuality of the beads.

Table 3. Single nucleotide primer extension assays for genotyping in the microar-ray format.

Assay feature Format References

Detection primer immobilised on the microarray

Radioactively labelled ddNTPs

(Pastinen et al. 1997)

On-slide temperature cy-cling

(Fortina et al. 2000)

“Array-of-arrays” (Lindroos et al. 2001)

Detection primer immobilised on the microarray, fluores-cently labelled ddNTPs for detection.

APEX (Kurg et al. 2000)

High-density array (Fan et al. 2000) Primer extension in solution followed by capturing of ex-tended primers using generic tag microarrays

Low density array (Hirschhorn et al. 2000)

Tag array minisequencing with four colour fluorescence de-tection

“Array-of-arrays” (Lindroos et al. 2002; Lovmar et al. 2003)

High throughput assay using MALDI-TOF detection

MassARRAY (Ross et al. 1998; Buetowet al. 2001)

Semi-automated SNPstream system employing the tag-microarray principle.

“Array-of-arrays” in 384-well microtiter plate for-mat

(Bell et al. 2002) Ge-nomeLab, Beckman Coulter

Genotyping in PCR colonies (Polonies) on microarrays

Acrylamide layer or mi-croarray surface

(Mitra and Church 1999; Mitra et al. 2003)

Allele specific primer extension A variant of the primer extension reaction principle is to use two allele specific primers complementary to the SNP alleles. In the case of a perfect match of the 3’-nucleotide of the primer with the SNP allele, an extension of the primer can take place (Figure 2B) (Wallace et al.1979). Extension can be performed with the four ddNTPs labelled with the same fluorophore in a single reaction, or dNTPs of which one is labelled can be used and allows for the incorporation of several la-belled nucleotide analogues leading to an increase in signal intensity. Increased specificity of the reaction has been achieved by analysing RNA templates in conjunction with reverse transcription reactions to extend the primers with labelled dNTPs in the presence of trehalose

22

(Pastinen et al. 2000). The inclusion of trehalose allows for a higher reaction temperature stabilising the enzyme and increasing the geno-type discrimination. This system has been used in a large scale survey of the prevalence of disease mutations in different geographical regions of Finland producing more than 64,000 genotypes (Pastinen et al.2001). Another approach for increasing the specificity of genotyping by allele specific extension is the inclusion of apyrase in the reaction. This is employed in the apyrase mediated allele specific primer exten-sion (AMASE) reaction. Apyrase degrades the nucleotides when the reaction kinetics are slow, as is the case for mismatch hybrids, thereby reducing the risk for incorrect genotyping (Ahmadian et al. 2001; O'Meara et al. 2002). The inclusion of locked nucleic acids (LNA) in the reaction could also be an alternative to increase specificity, this has been successfully employed for allele specific PCR (Latorra et al.2003). LNA are synthetic nucleic acid analogues with high affinity for DNA and RNA which due to its bicyclic chemical structure, increases hybrid melting temperature (Tm).

Ligation based assays In the oligonucleotide ligation assay (OLA) (Landegren et al. 1988), the two SNP alleles are detected using allele specific probes and a liga-tion probe that will hybridise to its target sequence (Figure 3A). In the case of a perfect match between the allele specific probe and the tar-get, the junction between the probes is closed by a DNA ligase with the ability to discriminate between a perfect and mismatch probe.

Ligation in a microarray format was first described for mutation de-tection in 1998 (Gunderson et al. 1998) where single stranded target DNA was ligated to an oligonucleotide immobilised on the array sur-face. Ligation on microarrays with one ligation probe immobilised has been described (Lizardi et al. 1998), or using stem loop ligation probes immobilised on the microarray surface (Broude et al. 2001). Other examples are performing the ligation reaction in solution followed by capturing of the products on microarrays (Gerry et al. 1999; Busti et al.2002; Delrio-Lafreniere et al. 2004) or on microparticles (Iannone et al. 2000) through hybridisation to tag or zipcode oligonucleotides.

The use of circularisable oligonucleotide ligation probes, padlock probes (Nilsson et al. 1994), have been used for SNP genotyping using tag oligonucleotides in a limited scale on microarrays (Banér et al.2003). Padlock probes have specific target recognition sequences in their 5’ and 3’ ends and a connecting sequence between the target spe-cific regions [Nilsson, 1994 #552]. When hybridised to its target se-quence the two ends of the probe are brought adjacent to each other, and the junction is ligated when there is a perfect match.

23

Figure 3. Reaction principles for genotyping of single nucleotide polymor-phisms using ligation or hybridisation. The detection of the A-allele of an A to G transition is illustrated. In the oligonucleotide ligation assay (OLA, panel A), a ligation probe and an allele-specific probe are used. In case of a perfect match between the allele specific probe and the target sequence, the junction between the two probes can be closed with a DNA ligase. For the hybridisa-tion reaction in panel B, two allele specific probes, with the SNP complemen-tary nucleotide located in the middle, are hybridised to the target sequence. The reaction conditions are optimised so that only the perfectly matched probe-target hybrid is stable and detectable.

Hybridisation based assays In hybridisation based assays, the destabilising effect of a single nucleo-tide mismatch between an oligonucleotide and its target is used to dis-tinguish between sequence variants (Figure 3B). The stability of the probe-target hybrid is affected by temperature, ionic strength, the se-quence surrounding the SNP position, and formation of secondary structures in the probe and target. Thorough optimisation of reaction conditions and primer design is necessary to be able to perform highly multiplexed hybridisation assays.

In the GeneChip® assay, the problem with low sensitivity and speci-ficity of multiplex ASO hybridisation has been circumvented by using tens of different probes for every SNP to be analysed (Hacia et al.1996; Hacia et al. 1998). Another possibility to increase the sensitivity involves analyses of the thermodynamics of duplex formation of match and mismatch hybrids over a temperature gradient on the microarray (Fotin et al. 1998). This approach has been extended to independent temperature controlled “islands” on the microarray surface (Kajiyamaet al. 2003). In the dynamic allele specific hybridisation (DASH) assay, performed on a macroarray nylon membrane, the allele specific probe is hybridised to the target at a low temperature allowing complete hy-bridisation of both match and mismatch probe. The temperature is then gradually increased resulting in the dissociation of the mismatch probe prior the release of the perfectly matched probe (Jobs et al.2003). The negative charge of nucleic acids allow for a guided trans-

24

portation of probes, and enhanced hybridisation when an electric field is applied over the microarray has also been demonstrated (Edman et al. 1997; Sosnowski et al. 1997). The use of peptide nucleic acid (PNA) as probes on the microarray may increase the hybridisation sta-bility (Weiler et al. 1997). PNA mimics DNA, but with its amide backbone it is stable to nucleases and proteases and acidic environ-ments. The binding of PNA to DNA is stronger than the binding of ASO probes, and thereby shorter probes can be used reducing the risk of secondary structures and increasing the effect of a mismatch nucleo-tide on the PNA-target hybrids (Griffin et al. 1997; Ross et al. 1997; Alsmadi et al. 2003; Brandt et al. 2003). In the same way, locked nu-cleic acids (LNA) can be used to improve hybridisation efficiency as exemplified by Orum etal. for detection of the Factor V Leiden muta-tion, a risk factor for venous thromboembolism (Orum et al. 1999). Incorporation of modified nucleotides in the target sequence may de-stabilise secondary structures and increase the interaction with the probe (Nguyen and Southern 2000).

Towards PCR free genotyping Highly accurate and simple methods for high throughput genotyping of SNPs are needed for whole genome based association studies and LD mapping of disease loci. The major bottleneck for highly multiplex SNP genotyping is the need for amplification of the target sequences by PCR to increase the copy number of the template of interest, and for reducing the complexity of the three billion base pairs of DNA in the human genome. The design of primers for multiplex PCR reactions requires extensive evaluation of the primers regarding unwanted inter-actions between primers of different pairs, as well as the variation in yield of the different fragments amplified together. Progress in devel-oping software allowing for multiplex design has been made (Vieux et al. 2002; Yuryev et al. 2002; Kaderali et al. 2003). Different primer modifications of the primers have also been suggested that allow for a more even yield of different fragments (Brownie et al. 1997; Broude et al. 2001).

Genotyping directly on genomic DNA, without PCR amplification, has been shown using the Invader assay (Third Wave Technologies) in a microsphere format (Rao et al. 2003). Hybridisation of a perfect match allele specific probe and an invader probe generates a tripartite structure that is recognised and cleaved by a nuclease generating a de-tectable signal. However, this system requires a large amount of ge-nomic DNA for reliable signal intensities thereby restricting its appli-cability.

New high throughput genotyping systems are emerging, in which the design of multiplex PCR reactions is circumvented by performing

25

PCR subsequent to the allele recognition One such example involves multiplex oligonucleotide ligation in the SNPlex™ Genotyping System (Applied Biosystems). The ligation of a perfectly matched allele spe-cific probe and a ligation probe is followed by amplification of the liga-tion products by PCR using a universal primer pair. A similar strategy is applied in the SNPWave™ technology for multiplex genotyping (van Eijk et al. 2004). Padlock probes are used to identify the SNP allele, and following ligation, the probes are selectively amplified in a multi-plex strategy based on the amplified fragment length polymorphism technology using a single set of PCR primers.

Another approach to reduce the complexity of the target sequences prior to large scale genotyping of SNPs is to perform a restriction en-zymatic cleavage of the genomic DNA. Amplification is then per-formed on a selected subset of the fractionated genome using universal primers in a process similar to the amplified fragment length polymor-phism method (Kennedy et al. 2003; Matsuzaki et al. 2004). Callow etal. have recently describe a similar technique, where adapter se-quences on the target sequence allow the formation of a circular struc-ture during digestion of linear fragments, thus enriching the sequence of interest (Callow et al. 2004). A single pair of amplification primers complementary to the adapter sequences is then used for multiplex amplification. These two techniques are restricted to the analysis of SNPs located within the restriction fragments, limiting the selection of markers.

A promising approach for increasing the specificity of genotyping SNPs directly in genomic DNA is to combine a primer extension reac-tion with a ligation step. Two examples of such methods are the BeadArray technology from Illumina (Oliphant et al. 2002), and the use of molecular inversion probes in the system from ParAllele BioSci-ence (Hardenbol et al. 2003). The ligation product is amplified using a single set of universal PCR primers, which allows for an even amplifi-cation of many fragments in a highly multiplex fashion, without the need for optimisation of the amplification reaction.

DetectionIn the beginning of the era of microarray based genotyping technology, radioactivity was used as detection method (Southern et al. 1992). Fluorescence is the most widely used strategy today in the microarray format. Both indirect and direct labelling is applied. The Affymetrix GeneChip® system is a well known example, where an indirect label-ling strategy is used. The target sequence is labelled with a biotin-conjugated nucleotide to which a fluorescent streptavidin-

26

phycoerythrin conjugate is bound (Chee et al. 1996; Hacia et al.1996).

Many fluorophores are available for direct labelling of nucleotides, followed by detection using fluorescence microscopes, flow cytometry, CCD cameras or fluorescence scanners. The use of a single fluorophore is suitable in assays where the allele recognition is by another event, such as allele specific primer extension (Pastinen et al. 1997). In this case the fluorescence label is only used for detecting if the primer ex-tension has occurred. In a multiplex genotyping assay, where the fluo-rescent label is the basis for genotype assignment, using a single label on all four nucleotides reduces the throughput of the assay due to re-quirement of four parallel reactions This is discussed in the section “Results and discussion” of my studies. Three different fluorophores is in principle sufficient to analyse all possible SNPs in a single reaction, if both DNA strands are utilised (Kwiatkowski et al. 1994; Hirschhornet al. 2000). However, the most efficient labelling approach for multi-plex SNP genotyping assays is to use four fluorophores, one for each nucleotide (Kurg et al. 2000; Lindroos et al. 2002; Lovmar et al. 2003). The use of multiple fluorophores requires that they have distinct non-overlapping emission spectra to avoid “cross-talk” between the labels during array scanning and signal analysis.

Mass as labelling principle combined with detection by MALDI-TOF mass spectrometry is widely used in SNP and proteomics assays. Detection by MALDI-TOF is fast and highly sensitive, has a wide dy-namic range, and yields favourable signal-to-noise ratios. Since the de-tection is based on mass, no separate labelling step is needed. Mass-labels offer the possibility for highly multiplexed assays. Discrimina-tion between single nucleotides as in primer extension reactions may not, however, be completely unambiguous for all SNPs due to the small differences in mass of the four nucleotides. To increase the reso-lution of primer extension in SNP genotyping, the mass difference be-tween the measured allele-specific products should be as large as pos-sible. In the GOOD assay (Sauer et al. 2000; Sauer et al. 2002), charge tagging of the primers and phosphotioate bridges in the backbone of the primer are used. Following the extension reaction, the primer is digested leaving only a few nucleotides, the charge tag and the ex-tended nucleotide for detection.

27

Microarray preparation Array surfaces Synthesis of oligonucleotides in situ or attachment of pre-synthesised oligonucleotides exist for numerous types of solid supports such as dextran (Gingeras et al. 1987), nylon membranes (Saiki et al. 1989), nylon beads (Van Ness et al. 1991), magnetic particles (Day et al.1991), polypropylene (Matson et al. 1994; Matson et al. 1995), polys-tyrene (Nikiforov and Rogers 1995) and polyacrylamide gel pads (Guschin et al. 1997; Mitra and Church 1999).

Glass is important and attractive for use as solid support in microar-ray based analyses because it is non-porous, mechanically and chemi-cally stable and has a low auto-fluorescence. Glass is also easy to deri-vatise for immobilisation of chemically modified oligonucleotides. Many variants for chemical modifications have been proposed for co-valent attachment of oligonucleotides on glass surfaces. Oligonucleo-tides with an amino group or a disulphide group in their 5’-end are often used for chemical coupling to various types of modified surfaces such as glass (Joos et al. 1997; Rogers et al. 1999) or beads (Walsh et al. 2001).

In situ synthesis of oligonucleotides In situ synthesis of oligonucleotides on a surface was first described in 1991 (Fodor et al. 1991; Pease et al. 1994) and is the approach for the manufacture of microarrays by Affymetrix, the pioneering manufac-turer of microarrays for use in gene expression profiling (Figure 4). This in situ synthesis system allows for production of microarrays with a high density of oligonucleotides. However, the process requires ex-pensive wafers for guiding the synthesis, and has been struggling with low yield of full length oligonucleotides owing to incomplete deprotec-tion of the nucleotides during synthesis (McGall et al. 1997).

The use of other photosensitive groups, such as photosensitive ni-trobenzyl (Serafinowski and Garland 2003) or phosporamidites (Beier and Hoheisel 2000), has been suggested to solve the problem of in-complete deprotection. To circumvent the need of expensive wafers for in situ synthesis, the nucleotides can be deposited one by one through inkjet dispensing according to a pre-programmed pattern (Hughes et al. 2001). Table 4 summarises some progress in the field of microarray preparation during the past ten years.

Primer extension assays require free 3’-ends for the incorporation of nucleotides. The in situ system described above is thereby not suitable since these oligonucleotides are synthesised in the 3’-5’-direction.

28

Figure 4. In situ synthesis of oligonucleotides on a microarray surface utilising photolithography. A mask is applied on the slide and light is used to selectively deprotect the photolabile linker covering the array surface through the holes in the mask. Thereafter a nucleotide with a protectable group is added and cou-pled to deprotected linkers. Another mask is then applied, the protecting groups removed and the next nucleotide added. This process is then repeated until the desired length of the oligonucleotides is reached. (Picture adapted from www.affymetrix.com).

However, synthesis of oligonucleotides in situ in a 5’-3’-direction has been shown but at a lower yield than 3’-5’-synthesis (Beier and Hoheisel 2002). Another proposed strategy is to perform the synthesis of the oligonucleotides in situ in a 3’-5’-direction and then invert the oligonucleotide on the surface, resulting in free 3’-ends (Kwiatkowskiet al. 1999). One benefit of this procedure would be intact 3’-ends of the oligonucleotides inverted, compared to the proposed 5’-3’-synthesis. This inversion technology is used by QUIAtech Biotechnol-ogy (www.quiatech.com).

Immobilisation of pre-synthesised oligonucleotides A more generally applicable procedure for the manufacturing of oli-gonucleotide microarrays is to deposit pre-synthesised oligonucleotides directly on the surface (Table 4). This is done through contact printing using either pins with a capillary slot that is filled with the oligonucleo-tide solution, the pin-hole system or inkjet dispensing using a piezo printing head similar to those used in ordinary laser printers. The inkjet technology is favourable because smaller spots are delivered with a high reproducibility, there is also less evaporation during the printing procedure since the system is closed compared to that of contact print-ing.

29

Table 4. Progress over time in oligonucleotide deposition on microarrays.

Strategy Reference

Light directed in situ synthesis using photolitho-graphic masks

(Fodor et al. 1991; Pease et al.1994)

Physical deposition of pre-synthesised oligonucleo-tides on microarray surface

(Southern et al. 1992; South-ern 1996)

Contact printing using split pins (Schena et al. 1995)

In situ synthesis by inkjet deposition wafers needed Blanchard 1996

Inversion of in situ synthesized oligonucleotides (Kwiatkowski et al. 1999)

In situ synthesis using maskless micromirror photo-lithography

(Singh-Gasson et al. 1999; Nuwaysir et al. 2002)

In situ synthesis by inkjet deposition of nucleotides drop by drop

(Blanchard et al. 1996; Hughes et al. 2001)

In situ synthesis of oligonucleotides in 5’-3’-direction

(Beier and Hoheisel 2002)

Maskless light directed in situ synthesis of oligonu-cleotides in 5’-3’ direction

(Albert et al. 2003)

Light activated in situ synthesis in a microfluidic system

(Baum et al. 2003)

Deposition of spotted long oligonucleotides as an alternative to in situ synthesised probes

(Barczak et al. 2003)

Fountain pens for high density printing of oligonu-cleotides on microarrays

(Reese et al. 2003)

Quantitative SNP analysis Methods that allow highly accurate and sensitive quantitative analysis are needed for analysis of genetic variation on the level of DNA and RNA. The polymerase chain reaction would appear to be optimal for use in quantitative analysis because of the sensitivity and specificity of the reaction. However, the amount of PCR product does not reflect the initial amount of template because the amplification is not expo-nential throughout the process. Therefore PCR-based quantification cannot be performed by direct measurement of the amount of PCR product obtained. Competitive PCR analysis which uses parallel ampli-fication of a known amount of an internal or external standard, allows for absolute quantification of the target sequence, assuming the target and standard are amplified by the same efficiency. Kinetic PCR is an alternative approach for quantitative PCR analysis, where the amount of target is monitored in the initial, exponential phase of the PCR am-plification. The 5’-exonuclease assay (“TaqMan”) is the most fre-

30

quently used approach for kinetic, quantitative PCR (Holland et al.1991; Livak et al. 1995). Both competitive and kinetic PCR are often applied for determination of gene expression levels or gene copy num-ber (Laan et al. 1995; Karttunen et al. 1996; Stenman et al. 1999).

Since the minisequencing method allows highly specific distinction between two SNP alleles, it can be used for determination of the ratio between two alleles also when they are present in other than the het-erozygote 1:1 ratio. This special case of competitive PCR analysis is useful for determination of SNP allele frequencies in pooled DNA samples from a population (Olsson et al. 2000) or from affected and non-affected individuals in association studies (Arnheim et al. 1985). It has also been used for determination of the level of heteroplasmy for mitochondrial mutations (Suomalainen et al. 1993; Lagerstrom-Fermeret al. 2001) or chimerism between cells from the donor and recipient after stem cell transplantation (Hochberg et al. 2003). Quantitative analysis of SNPs on the level of RNA allows identification of imbal-anced expression of two alleles of the same gene (Yan et al. 2002; Pastinen et al. 2004) (Study V). Multiplexed methods allowing analysis of allelic differences in gene expression (Lo et al. 2003), would be use-ful for scaling up studies targeted at individual genes to a genome wide scale.

31

PRESENT STUDY

Introduction The studies included in this thesis describe the development of mi-croarray based minisequencing with fluorescence detection for geno-typing of SNPs, and its application in pharmacogenetics. The studies were initiated at an early stage of microarray technology. At that time only a limited amount of SNP information of low quality was available in the databases. In parallel with the completion of the sequence of the human genome, the information on SNPs in public databases has in-creased substantially and the quality of the information has improved. Thus a much more efficient selection of SNPs to serve as genetic markers is feasible today. In the present study different microarray surfaces were evaluated to identify the best one for the minisequencing system. Improvements in technology for array spotting have increased the reproducibility of our SNP assays. Better availability of fluores-cently labelled nucleotide analogues and instruments for scanning mul-tiple fluorophores has allowed an increase in throughput of the mi-croarray system during the progress of my thesis work.

AimsThe specific aims of my thesis work have been

to develop a microarray minisequencing system with fluores-cence detection for high throughput genotyping of SNPs. to establish the microarray minisequencing system for quanti-tative analysis of SNPs in DNA and RNA. to apply the microarray minisequencing system for identifica-tion of SNPs predictive of response to antihypertensive treat-ment.

32

Materials and Methods Study samples DNA samples from healthy Swedish volunteer blood donors were used in Study I. A pooled DNA sample comprising equal amounts of DNA from 100 Swedish volunteer blood donors was used to determine the allele frequency of the SNPs in Study II and IV. In Study II and IIIa-c, DNA samples from participants of the multicenter Swedish Irbesartan Left Ventricular Hypertrophy Investigation versus Atenolol (SILVHIA) trial (Malmqvist et al. 2001) were analysed. DNA samples collected from patients with different types of cancers who had under-gone stem cell transplantation during the period 1994-2001 at Uppsala University Hospital were also analysed in Study IV. Blood or bone marrow samples had been collected from the stem cell donor and the recipient before the transplantation and from the recipient at different time points ranging from one month to eight years during the follow up after the transplantation. In Study V, the samples for the quantifi-cation standard curves were prepared by mixing DNA of different genotypes for the SNPs analysed to get a series of samples with differ-ent proportions of the SNP alleles. In addition, DNA and RNA sam-ples from Human Umbilical Vein Endothelial Cells (HUVEC) and Human Aortic Endothelial Cells (HAEC) from Cascade Biologics, Inc. (Portland, OR, USA) were used in Study V.

The DNA in Study I, II, and V was extracted using the Wizard Ge-nomic DNA Purification Kit (Promega, Madison, WI, USA) or QiAamp DNA Blood Mini Kit (Qiagen, Hilden, Germany), while a standard salting-out method was used for the samples in Study IV (Miller et al. 1988).

The SILVHIA trial This prospective multicenter trial was originally designed to evaluate the efficacy of two antihypertensive drugs on the change in left ven-tricular mass (LVM) in hypertensive individuals. A total of 115 men and women with primary mild to moderate hypertension and LVH were randomised to receive 150 mg of the angiotensin II type 1 recep-tor blocker irbesartan or 50 mg of the 1-receptor blocker atenolol once daily in a double blind parallel group trial. For 66% of the sub-jects receiving irbesartan and 39% receiving atenolol the dose was dou-bled after six weeks of treatment due to a diastolic blood pressure ex-ceeding 90 mmHg. Fourteen patients discontinued the trial. DNA and echocardiography data was available from 97 subjects and these were analysed in Studies II, IIIa, IIIb and IIIc.

33

Blood pressure was measured by trained nurses, using a mercury sphygmomanometer after the patients had rested for at least 10 min-utes in seated position, and reported as an average of three measure-ments taken one minute apart. LVH was considered to be present with a left ventricular mass index (LVMI) of > 131 g/m2 for men and > 100 g/m2 for women.

The blood pressure responses presented in Studies II, IIIa and IIIb, and the change in LVMI presented in Study IIIc relates to twelve weeks of monotherapy.

SNP selection and primer design The SNPs analysed were primarily selected from the public SNP data-bases dbSNP (http://www.ncbi.nlm.nih.gov/SNP/) and TSC (http://snp.cshl.org/) databases. The SNPs for Study II were located in candidate genes in the major pathways involved in blood pressure regulation. SNPs that had been shown or suggested to be important markers for blood pressure regulation in previous studies were identi-fied from the literature. SNPs likely to be informative due to their fre-quency and genomic location were selected from the databases. In total 98 SNPs in 25 genes were included in the Study. The SNPs used in the marker panel in Study IV were selected from the dbSNP database. Selection was based on their genomic location, presence of allele fre-quency data for the Caucasian population and validation status. A few SNPs from Study II were also included in this panel. For Study V, SNPs located in coding parts of genes were analysed. Ten SNPs were selected from the panel analysed in Study II and from the dbSNP data-base.

For the multiplex PCR reactions, primers were designed to have similar Tm-values and two or three A- or C-nucleotides in their 3’-end to avoid extendable primer-dimers during the amplification. Minise-quencing primers were designed for analysis of the SNPs in both DNA polarities. The length of the primers was adjusted to reach similar Tm,with a range of 18-25 nucleotides. The minisequencing primers to be immobilised on the microarray surface in Studies I, II and V had a spacer sequence of 15 T-residues and a 5’-amino group. The minise-quencing primers for the cyclic minisequencing reactions in Studies IV and V contained a 5’ tag-sequence from the GenFlex™ Tag collection (Affymetrix, Santa Clara, CA, USA). Complementary tag sequences, to be immobilised on the microarrays, contained a 15 T-spacer and an amino group in their 3’-end.

All primers were designed using the Oligo Primer Analysis software v6.65 (Molecular Biology Insights Inc., Cascade, CO, USA) or Primer3 (Rozen and Skaletsky 2000), except for the minisequencing primers

34

used with the SNPstream system in Study V, for which the Auto-primer software (Beckman Coulter Inc., Fullerton, CA, USA) was used.

Polymerase chain reaction All PCR reactions were performed in a Programmable Thermal Con-troller (PTC225, MJ Research, Watertown, MA, USA) with the ther-mostable AmpliTaq® Gold DNA polymerase. The PCR fragments in the multiplex reactions were of different length, enabling verification of the success of the amplification by agarose or polyacrylamide gel electrophoresis.

The PCR reactions were optimised with respect to MgCl2 concen-tration, amount of DNA polymerase and number of amplification cy-cles. Primer concentrations were adjusted to achieve an even and effi-cient amplification of all fragments. In Study II, the amplification was performed using a touchdown procedure (Don et al. 1991) instead of the three steps of denaturation, annealing and extension used in the other studies. In addition, to uniform the amplification of multiple sequences, universal tail sequences were included in the 5’-ends of the primers (Shuber et al. 1995). These tail sequences were used in Studies II and IV to introduce a biotin residue in a second amplification for immobilisation of the product in the solid phase minisequencing reac-tions using streptavidin-coated microtiter plate wells.

Preparation of microarrays In order to analyse many samples in parallel on each microarray slide an “array-of-array” format was used in all studies (Figure 5). In each of the “subarrays” the same pattern of oligonucleotides is printed allowing parallel analysis of many samples for the same set of SNPs on one sin-gle microarray. This “array-of-arrays” format, first described by Pasti-nen etal. (Pastinen et al. 2000), has been adopted to different minise-quencing systems (Lindroos et al. 2001; Lindroos et al. 2002), as well as for genotyping using padlock probes (Banér et al. 2003). The sepa-rate reaction wells are formed by positioning a silicone grid on the mi-croarray (Figure 5). The silicone grid is cast by pouring polydimethylsi-loxane (Elastosil® 601 or 625, Wacker-Chemie GmbH, Munich, Ger-many) in an inverted 384- or 96-well microtiter plate, resulting in cone shaped reaction chambers.

35

Figure 5. The “array-of-arrays” format (A) allows for parallel analysis of 80 samples for 200 SNP markers resulting in 16,000 genotypes (left feature, used in Studies I, II and V) or of 14 samples for 500 SNP markers resulting in 7,000 genotypes (right feature, used in study IV). B: The custom made mi-croarray reaction rack can hold up to three microarray slides. The rack is made of aluminium and placed on a heat source during the incubation. The reusable silicone grid is used to separate the different samples on the microar-ray.

The oligonucleotide microarrays used in my studies were all printed in-house. In Study I, a custom built modified industrial robot (Isel, Eiter-field, Germany) controlled by an MCM-310 operating system and NUMO 6.0 software (Merval, Pietasaari, Finland) equipped with two CMP2 pins (TeleChem International Inc., Sunnyvale, CA, USA) was used. In Study II, IIIa-c, IV and V, the oligonucleotides were printed using a ProSys 5510A instrument (Cartesian Technologies Inc., Irvine, CA, USA) equipped with one (Study V) or four stealth microspotting pins (SMP3 or SMP3B, TeleChem International Inc., Sunnyvale, CA, USA). The spots had a diameter of approximately 150 µm, and a cen-tre-to-centre distance of 200-250 µm.

In Studies II, IV and V the oligonucleotides were immobilised on CodeLink™ activated slides (Amersham Biosciences, Uppsala, Sweden) after they had been identified as most suitable for the microarray based minisequencing in Study I

Minisequencing in the microarray format In Study I, we developed a microarray system with the minisequencing primers immobilised on the surface and extension with fluorescently labelled ddNTPs (denoted Method I, Figure 6). Method I was further used in Studies II, IIIa-c and V. Multiplex PCR products were com-bined and concentrated by ethanol precipitation (Studies I and II) or using Microcon YM-30 Centrifugal Filter Devices (Millipore Corpora-tion, Bedford, MA, USA) (Study V).

The PCR products were allowed to anneal to the minisequencing primers for 40 min at 37°C in a humid chamber, followed by rinsing and reassembling the rack. In Studies I and II the reaction mixture con-tained one Tamra-labelled ddNTP (Perkin Elmer Life Sciences, Boston,

36

MA, USA) and the three other ddNTPs unlabelled together with Dy-naSeq DNA polymerase (gift from Finnzymes OY, Helsinki, Finland) or Thermo Sequenase™ DNA polymerase (Amersham Biosciences, Uppsala, Sweden). In Study IV and V, the ddNTPs were labelled with four different fluorophores (Texas Red-ddATP, Tamra-ddCTP, R110-ddGTP and Cy5-ddUTP) and used together with Thermo Sequenase™

DNA polymerase (Amersham Biosciences, Uppsala, Sweden). The extension reactions were performed at 68°C for 5 min or 55°C for 15 min.

In Study IV and V, a tag array based minisequencing system was utilised (denoted Method II) (Figure 6). Excess primers and dNTPs from the PCR were removed by treatment with exonuclease I and shrimp alkaline phosphatase (USB Corporation, Cleveland, OH, USA). The extension reactions were performed in solution with the tagged minisequencing primers, fluorescently labelled ddNTPs (Texas Red-ddATP, Tamra-ddCTP, R110-ddGTP and Cy5-ddUTP, Perkin Elmer Life Sciences, Boston, MA, USA) and Thermo Sequense™ DNA polymerase for 55 or 99 cycles of 95°C and 55°C for 20 sec each. The extended primers were hybridised to oligonucleotides complementary to the tag sequences, immobilised on the microarrays, for 2.5 hours at 42°C followed by washing of the slide.

Figure 6. Minisequencing on microarrays with four-colour fluorescence detec-tion. Detection of a heterozygous SNP is illustrated. In Method I (panel A) minisequencing primers are immobilised on the microarray surface. The SNP-containing fragments amplified by PCR are annealed to the primers and ex-tended with fluorescently labelled ddNTPs in situ. In the tag array based minisequencing system (Method II, panel B) cyclic primer extension reactions are performed in solution with fluorescently labelled ddNTPs. The primers are then captured by hybridisation to complementary tag sequences (gray) immobilised on the microarray surface.

37

In Study V, the semi-automated SNPstream system (GenomeLab™,Beckman Coulter Inc., Fullerton, CA, USA) was used (denoted Method III). The removal of PCR reagents and the extension reactions were performed as for Method II, with the exception that the reagents used were part of the SNPware reagent kit. The SNPs to be analysed were pooled according to nucleotide variation, and extension reactions were performed with one Tamra-labelled ddNTP and one Fluorescein-labelled ddNTP for 45 cycles of 94°C and 55°C for 30 sec each.

Minisequencing in the microtiter plate format The biotinylated PCR products were immobilised in a microtiter plate coated with streptavidin (Combiplate 8, Labsystems, Helsinki, Finland) and the unbiotinylated strand was removed with alkali treat-ment (Syvanen et al. 1993). The minisequencing mixture, containing the appropriate tritium labelled dNTP (Amersham Biosciences, Upp-sala, Sweden), AmpliTaq® DNA polymerase (Applied Biosystems, Foster City, CA, USA) and the minisequencing primer was added. The extension reaction was allowed to proceed for 10 min at 55°C. The extended primers were released with alkali and the amount of incorpo-rated tritium labelled nucleotide was measured.

Signal detection The fluorescence from the incorporated ddNTPs was measured using a single colour confocal microarray scanner, ScanArray 3000® instru-ment in Study I. This system had been upgraded to four colour detec-tion in the ScanArray 5000® instrument used in Study II and IV, and to ScanArray® Express in Study V (Perkin Elmer Life Sciences, Boston, MA, USA). The excitation laser in the one colour scanner was Green HeNe 543.8 nm, while Blue Argon 488 nm, Yellow HeNe 594 nm and Red HeNe 632.8 nm were also available in the four colour array scan-ner. In Studies IV and V, reaction control spots were used to adjust the photomultiplier tube gain to obtain equal signal intensities for all four spectra. The fluorescence signals were extracted using the QuantAr-ray® analysis software.

The fluorescence from the ddNTPs in the SNPstream assays (Study V) was measured with a CCD camera with excitation lasers of 488 nm and 532 nm integrated in the genotyping system. The signal intensities were extracted using the SNPstream Image™ and GetGenos™ software.

For minisequencing in the microtiter plate format, the amount of incorporated tritium labelled dNTPs was measured using a WinSpec-tral™ 1414 liquid scintillation counter (Wallac, Turku, Finland).

38

Genotype assignment The fluorescence signals were corrected for the background in each “subarray” separately. Genotypes were assigned to the samples by cal-culating a ratio between the fluorescence signals for one allele over the signal of the other allele, or by calculating a signal intensity fraction by dividing the signal from one allele with the sum of the signals from both possible alleles.

Statistical analysisIn Study II and IV, Hardy-Weinberg equilibrium was analysed by comparing the observed and expected genotype frequencies using a 2-test with one degree of freedom. In Study II, the most probable haplo-types were assigned using the expectation maximisation algorithm in the Haplo program (http://info.med.yale.edu/genetics/kkidd) (Hawley and Kidd 1995) and the subset of SNPs required for distinguishing between haplotypes (Johnson et al. 2001) were identified by visual inspection. Analysis of variance (ANOVA) was used to determine the relationship between SNP genotype and change in blood pressure or LVM upon antihypertensive treatment in Studies II and IIIa-c. In Study IIIb, correction for multiple testing was performed by a permu-tation test (Churchill and Doerge 1994). Multiple stepwise regression analysis was used in Studies II and IIIc to identify combinations of SNP genotypes in relation to the change in blood pressure and LVM related to the response to the antihypertensive treatment. In Study V, two-sample t-test with two-tailed significance levels were performed to determine the lowest level of detection of a specific allele for the quan-tification standard curves and to evaluate the imbalanced expression of the two alleles of the SNPs in the cell lines.

Results and discussion Microarray based genotyping Immobilisation of oligonucleotides The success of a genotyping assay in the microarray format relies on the reaction principle and the surface of the microarray support. In Study I we evaluated different surfaces for covalent attachment of oli-gonucleotides on a glass surface. Four commercially available slides, coated with different chemically reactive groups, and one uncoated slide were compared with slides coated in-house with aminopropyl-triethoxysilane and 1,4-phenylene diisothiocyanate (PDC) (Guo et al.

39

1994). The different types of slides and their reactive groups are listed in Table 5.

The six slide types were compared with respect to their capacity to bind oligonucleotides, background fluorescence before and after the minisequencing reaction, and for their performance in the microarray based minisequencing assay for SNP genotyping. The results were re-lated to the in-house coated isothiocyanate slide that served as refer-ence.

As can be seen in Table 5, the 3D-Link™ slides, coated with a three dimensional hydrophilic polymer attaching amino-modified oligonu-cleotides, showed an almost eight times higher binding capacity rela-tive to the isothiocyanate slide. The mercaptosilane coated slide, bind-ing disulfide-modified oligonucleotides, showed a two times higher binding capacity. The EZ-RAYS™ slides, also with a three dimensional surface, only bound one fifth the amount of oligonucleotide compared to the isothiocyanate slide (Table 5).

Table 5. Comparison of immobilisation chemistries

Slide name(Manufacturer)

5’-modification of primer Reactive group on the slide

Relative bind-ing capacityc

Isothiocyanate (In-house)

NH2-(CH2)6-O-PO2-oligonucleotide

-N=C=S 1.0

‘SuperAldehyde’ (Tel-eChem International)

NH2-(CH2)6-O-PO2-oligonucleotide

-CHO 0.090/0.20

3D-Link™

(SurModics Inc.)aNH2-(CH2)6-O-PO2-oligonucleotide

Hydrophilic polymer with amine-reactive groups

7.7

Mercaptosilane HO-(CH2)6-S-S-(CH2)6-O-(PO)2-oligonucleotide

H2C=CH-CO-NH-R-oligonucleotide

-SH 1.9

EZ-RAYS™

(Mosaic Technologies)bHO-(CH2)6-S-S-(CH2)6-O-(PO)2-oligonucleotide

H2C=CH-CO-NH-R-oligonucleotide

Three dimen-sional with SH-groups

0.19

Unmodified (OCH3)3-Si-R-oligonucleotide

None 0.80

a Slides now called CodeLink™ activated slides (distributed by Amersham Biosciences). b Company now named Matrix Technologies. c The 3’-end of the immobilised oligonucleotides were labelled with a fluorescent dNTP using terminal deoxynucleotidyl transferase. The relative binding capacity corre-sponds to the relative signal intensity compared to the isothiocyanate slide.

40

After performing minisequencing reactions with Tamra labelled ddNTPs, the fluorescence was measured outside the spots in each “su-barray”, to determine the background signal intensity caused by unspe-cific binding of the ddNTPs to the surface (Figure 7). The uncoated slide showed the lowest background fluorescence, while the 3D-Link™

and the mercaptosilane slides had the highest background signals. This resulted in signal to noise ratios that were only slightly better for the 3D-Link™ and mercaptosilane slides although the signal intensities from the oligonucleotides were up to ten times higher than those from the isothiocyanate slide in the minisequencing assay.

Figure 7. Background fluorescence signals measured after the minisequencing reactions. SuperAldehyde 1 and 2 respectively, correspond to two different protocols for processing the slides after the printing of the oligonucleotides. The background fluorescence is given as arbitrary fluorescence units.

In addition to a high binding capacity and low background fluores-cence, the ability to unequivocally discriminate the heterozygote from the homozygotes is important for a successful SNP genotyping. Minis-equencing reactions were performed on the 3D-Link™ and mercaptosi-lane slides, i.e. the slides with the highest oligonucleotide binding ca-pacity and signal intensities compared to the other slides evaluated. The genotypes were determined without conflict for both slide types. For the heterozygote genotype, the minisequencing reaction on the mercaptosilane slides resulted in better signal intensity ratios and thereby the best genotype discrimination (Figure 8). The lower signal intensity ratios for the 3D-Link™ slides could be explained by the higher background fluorescence. However, the slightly lower signal intensity ratios, did not affect the genotype assignment of the SNPs from the minisequencing reaction.

41

Figure 8. The genotype discrimination was evaluated after minisequencing reactions using both polarities (denoted co and nc) for three SNPs on the 3D-Link™ and mercaptosilane slides. A ratio between the fluorescence signal in-tensities for the two alleles for each SNP was used to assign the genotype. The numbers indicate the signal intensity ratios higher than 10. The correct geno-type is indicated after the SNP name to the left of the diagram.

Based on the results from this evaluation of different slides for covalent attachment of oligonucleotides, the mercaptosilane slides would be the slide of choice. As these are not commercially available, the 3D-Link™

slides became the slide of choice for our microarray based minise-quencing reactions and have been used since. This type of slides has also been demonstrated to be highly sensitive and specific in expression profiling assays (Ramakrishnan et al. 2002).

Genotyping with single-colour fluorescence detection The microarray minisequencing system used in Study I was also used in a large scale genotyping study (Study II). We genotyped 74 SNPs in 97 hypertensive individuals from the SILVHIA trial (see also section “Applications“). Minisequencing primers were used for one polarity of the DNA. Due to variability in the efficiency of the minisequencing reaction, the other polarity had to be included for approximately 20% of the SNPs to be able to assign a genotype. To further validate the results of the 7200 genotypes generated in the microarray based assay, 200 randomly selected genotypes were reanalysed in the microtiter plate based minisequencing assay. Fully concordant results were ob-tained, thus providing evidence for the high accuracy of the microarray based genotyping system.

In Study II the signal intensities were manually evaluated to exclude weak signals, followed by calculating signal intensity fractions to assign

42

the genotypes. This part of the work can be facilitated by using cluster analysis for visualisation of the genotyping results. For each SNP, scat-ter plots with the sum of the fluorescence signals for the two alleles plotted against the signal intensity fraction plotted on the x-axis are generated. This should theoretically result in three distinct clusters of signal intensity fractions close to 0, 0.5 and 1, respectively for the three genotypes. The generation of such scatter plots can be done by pro-gramming Excel Macros or utilising the SNPSnapper software devel-oped by Juha Saharinen (National Public Health Institute, Helsinki, Finland).

By calculation of the average of the signal intensity fractions for the separate genotypes for each SNP, the variation in genotype discrimina-tory power was evaluated for the SNPs analysed in Study II (Figure 9). The three clusters for each SNP are separated with an average distance of 0.32. However, the variability between SNPs is high and the reason for this is still to be elucidated. It is likely a combination of different factors such as background fluorescence and signal intensities. Due to unequal efficiency of incorporation of ddNTPs by the DNA poly-merase the signal intensity fractions may be skewed depending on the nucleotide variation of the SNP (Lindroos et al. 2002). The sequence surrounding the SNP position is also suggested to affect the annealing of the primer to the template and thereby the efficiency of the exten-sion (Norton et al. 2002). For SNPs with small differences between the signal intensity fraction for homozygous and heterozygous genotypes, analysis of the strand of the opposite polarity can be an alternative for better allele discrimination.

Figure 9. Power of genotype discrimination. The SNPs analysed are given be-low the panel. The power of genotype discrimination is defined as the dis-tance between the homozygous genotype clusters for allele 1 (triangle) or allele 2 (square) to the heterozygote (diamond). This distance is given as the absolute value when subtracting the average signal intensity fraction for the homozygous samples from the average signal intensity fraction for the het-erozygous samples.

43

Genotyping with four-colour fluorescence detection In Study IV and V, Method I and Method II (Figure 6) were used with four-colour fluorescence detection. The four colour system was possi-ble when a microarray scanner with multiple excitation lasers and nu-cleotide analogues labelled with different fluorophores became avail-able. The four fluorophores were selected to have distinct, non-overlapping emission spectra, to minimise the level of cross-talk be-tween the different detection channels. Compared to Study I, perform-ing the assay with four fluorophores increases the problem with un-equal efficiency of incorporation of the four nucleotides. Both the dif-ferences between the nucleotides and the different chemical structure of the fluorophore affects the efficiency of the DNA polymerase (Lindroos et al. 2002). Some fluorophores are also more prone to photo-bleaching during the scanning process, resulting in differences in signal intensities.

One major benefit of the four-colour system is that each sample is assayed for all four nucleotides in one single “subarray” on the microar-ray. This circumvents the variability between spots of different “subar-rays” caused by differences in the performance of the pins during print-ing and irregularities in the microarray surface. This also allows for an increased number of samples that can be analysed in parallel on each slide, from 16 to maximally 80 (Figure 5), which also facilitate the data analysis using cluster analysis.

Accuracy and sensitivity of allele quantification In Study V, genotyping of dilution series of mixed samples with differ-ent allelic proportions was performed to determine the accuracy of the two microarray based minisequencing systems Methods I and II (Figure 6). The accuracy was determined using the coefficient of determina-tion (R2) that describes how well a regression line fits the data points of the standard curve. For Method I, six of the SNPs showed R2 0.95 for the regression lines while eight SNPs showed R2 0.95 for Method II (Table 6). The overall results for the ten SNPs for the two methods show that accurate quantification would be possible regardless of which method used, even though variation in accuracy is seen between SNPs as well as between the two genotyping systems.

The sensitivity of detecting a minority allele in a mixed sample was determined for Method I and II using a two-sample t-test. The detec-tion limit was defined as the percentage of the minority allele in the mixed sample, for which the signal intensity ratio differed from the signal intensity ratio in the corresponding homozygous sample with a p-value < 0.05.

44

Table 6. Results for the regression lines describing the accuracy and limit of detec-tion for the analysis of mixed samples of a dilution series.

SNP Method R2a Sensitivity range (%)b

p >0.05 p < 0.05

rs4331 ACE I 0.95 4.9 7.5

II 0.97 2.1 4.9

rs1042713 ADRB2 I 0.99 2.1 4.9

II 0.98 0 2.1

rs1042714 ADRB2 I 0.96 8.9 14*

II 0.98 5.8 8.9*

rs1042718 ADRB2 I 0.99 8.9 14*

II 0.99 3.8 5.8*

rs1042719 ADRB2 I 0.90 3.8 5.8

II 0.96 0 1.0

rs1799983 ECNOS I 0.99 5.8 8.9*

II 0.97 2.5 3.8*

rs5351 EDNRB I 0.87 3.8 5.8

II 1.0 0 1.0

rs5925 LDLR I 1.0 2.1 4.9

II 0.98 2.1 4.9

rs5930 LDLR I 0.94 5.8 8.9

II 0.63 21 33

rs1433099 LDLR I 0.75 ND ND

II 0.88 3.8 5.8 a Coefficient of determination describing the fit of the regression line to the data points of the standard curves. b Allelic range within which the minority SNP allele can be detected. The higher and lower percentages correspond to the mixture with a signal intensity ratio significantly different (p < 0.05) or indistinguishable (p > 0.05) from the signal intensity ratio of the corresponding homozygous sample or heterozygous sample (*). ND: not possible to determine due to the scatter of the data points.

For Method I, between 5% and 9% of the minor allele was detectable for seven of the SNPs analysed, while Method II was slightly more sen-sitive detecting between 2.5% and 6% (Table 6). These results agree well with previous results of allele frequency determination in pooled DNA samples using Method II (Lindroos et al. 2002), as well as for

45

other primer extension assays performed for individual SNPs (Mohlkeet al. 2002; Norton et al. 2002; Hochberg et al. 2003). A more efficient extension of the primers in the cyclic reactions could explain the dif-ferences seen between Method I and II. This as a result of greater in-teraction of the primers and template sequences due to less steric hin-drance.

The sensitivity of allele determination of the microarray based sys-tem is slightly lower than for the microtiter plate system of the minis-equencing method. In Study IV we determined the detection limit to 1% using the microtiter based minisequencing system with tritium labelled nucleotides. Standard curves were constructed for three SNPs after genotyping of a series of mixed samples with different allelic ra-tio. Excellent quantitative performance was seen with coefficients of determination (R2) of 0.99 (Figure 10).

Figure 10. Standard curves displaying the accuracy and sensitivity of quantita-tive genotyping of three SNPs using the minisequencing microtiter plate sys-tem. For the SNP rs13841 (panel A) the allelic range is from a homozygous sample to a heterozygous sample; for the SNP rs714195 (panel B) from a heterozygous sample to a homozygous sample; and from a sample homozy-gous for allele 1 to a sample homozygous for allele 2 for the SNP rs714195 (panel C). The mean cpmallele1/cpmallele2 ratio for three parallel minisequenc-ing reactions is plotted as a function of the known allelic ratio in the mixed samples. The corresponding percentage of allele 1 is indicated above the x-axis. The arrows indicate the cpmallele1/cpmallele2 ratio for the two samples of the dilution series analysed individually.

These results are in good accordance with previously studies on deter-mination of SNP alleles in pooled samples (Olsson et al. 2000) and the level of heteroplasmy in mitochondrial DNA (Lagerstrom-Fermer et al.2001) using this system. The higher sensitivity of the microtiter based system compared to the microarray based systems could be explained by the use of unmodified nucleotides or the stringent washing during the reaction that removes everything but the DNA strand immobilised in the microtiter plate well, thus reducing the complexity of the reac-tion.

46

Detection of alleles in mixed DNA samples In Study IV a two phase minisequencing strategy for the study of chi-merism was developed as a quantitative tool in the follow-up of stem cell transplantation in leukaemia patients. A panel consisting of 51 SNPs located on different chromosomes was assembled. The SNPs were selected to be located far apart to reduce the possibility of LD between them, and thus make them useful as markers. In a first step, stem cell donor-recipient pairs were genotyped for the 51 SNPs using the tag array based minisequencing system (Method II, Figure 6). In total, 25 donor-recipient pairs, of which nine donors were unrelated to the recipient and sixteen were siblings to the recipient, were screened in order to identify informative markers differing in genotype between the two individuals. Between eight and nineteen SNPs (average twelve) were identified for the unrelated pairs while four to ten (aver-age six) were identified for the related pairs which is in accordance with Monte Carlo simulated data based on the population frequencies of the SNPs analysed.

For nine of the donor-recipient pairs, samples taken prior to stem cell transplantation and follow-up samples from different time points ranging from one month to eight years after transplantation were avail-able. The development of chimerism in these samples were analysed in the second step of the minisequencing strategy, where a sensitive and quantitative analysis of the informative SNPs was performed using minisequencing in a microtiter plate format with tritium detection (Syvanen et al. 1993). With prior knowledge and as shown in section “Accuracy and sensitivity of allele quantification”, this method is well suited to accurately determine low amounts of an allele in a mixed sample. In figure 11, the analysis of the development of chimerism after stem cell transplantation is illustrated.

Figure 11. Examples of the development of chimerism using SNPs differing in genotype between the donor and the recipient. The follow-up samples were analysed using microtiter plate based minisequencing. Stem cell transplanta-tion was performed at time point zero, and the fraction of allele 1 of the in-formative SNP, provided by the donor, was followed in the recipient. In case 13, the two informative SNPs were of different homozygous genotypes in the donor and recipient, while one of the SNPs for case 15 was heterozygous lead-ing to a range in allele fraction from 0.5 to 1.

47

In the future, as the microarray based minisequencing systems will provide a more sensitive analysis for allele determination, the study of chimerism will be facilitated because it will be possible to screen for informative SNPs and perform the quantitative analysis in one reac-tion.

Detection of allelic imbalance In Study V we evaluated the two microarray based genotyping systems (Methods I, II) for quantitative SNP analysis on the level of RNA to study allelic differences in gene expression. In addition, a prototype version of the semi-automated SNPstream system (Bell et al. 2002) (GenomeLab, Beckman Coulter) that utilises the tag array principle, as in Method II, in a 384-well microtiter plate format (Method III) was included in the analysis.

Ten SNPs were analysed in two endothelial cell lines in this pilot project. The SNPs were first genotyped in genomic DNA (gDNA) of the cells in order to identify heterozygous markers that could be used to asses the level of expression of the two alleles in RNA. Three SNPs, rs4331 ACE, rs1042719 ADRB2 and rs5351 EDNRB, were heterozy-gous in the human umbilical vein cells (HUVEC) and three SNPs in the LDLR gene (rs5925, rs5930, rs1433099) were heterozygous in the human aortic cells (HAEC). These SNPs were further analysed in cDNA (RNA) from the cells, in parallel with the gDNA in five repli-cate assays for the three methods (Figure 12).

Figure 12. Analysis of imbalanced expression of alleles for six SNPs in two cell lines using Methods I, II and III. Imbalanced expression of alleles was deter-mined by comparing the signal intensity ratio of the amounts of labelled nu-cleotide incorporated for the two SNP alleles for the cDNA, with the corre-sponding ratio of the gDNA sample where the two alleles are present in a 1:1 ratio. The allele ratios represent mean values from five parallel reactions with one standard deviation.

48

This comparison gives the relative expression levels of the two alleles. To determine the absolute level of expression requires the use of inter-nal standards expressed at constant levels. We identified 1.4-1.5-fold differences in allelic expression for four of the six analysed SNPs (Fig-ure 6). Method III showed similar results as Methods I and II, although there are several differences in the reaction conditions. Methods I and II were performed with four colour label of the ddNTPs, and with syn-thetic sequences for controlling the different steps of the reactions. Since Method III is a commercial system, the reagents are provided in a kit, making adjustments of the reaction conditions to be more similar to Methods I and II impossible. The assay design for Method III in-volves two fluorophores on the nucleotides, and detection with a CCD camera instead of a confocal microarray scanner. These differences hampered the comparison of the three systems, but the results show the possibility of using the SNPstream system for detecting imbalance expression of alleles at a high throughput scale.

The results of Method III are promising for future large scale screen-ings for imbalanced expression of alleles in the search for regulatory variants in the untranslated regions of the genes. Such genetic variation has been suggested as a common cause of both normal and disease re-lated inter-individual variation in complex phenotypes (Hudson 2003).

SNP selection In Studies II and IV, the microtiter plate format of the assay was used to determine the allele frequency of the SNPs in the Swedish popula-tion. A pooled DNA sample with equal amounts of DNA from 100 healthy volunteer blood donors was analysed and SNPs with a minor allele frequency of 1% were excluded from the studies. These SNPs would not be informative due to the limited number of samples avail-able in the studies.

Of the 98 SNPs selected for Study II, sixteen SNPs (16%) were ex-cluded from further analysis due to a minor allele frequency of less than 1%. The allele frequency for the remaining SNPs compared well with the frequency data for European populations available in the da-tabases. The number of excluded SNPs would probably have been lower if the SNP selection had been performed today, when a larger fraction of the SNPs are validated with allele frequencies available for several populations as a result of the sequencing of the human genome (Lander et al. 2001; Venter et al. 2001) and the HapMap project (HapMapConsortium 2003). This was confirmed by the results from Study IV where SNP selection was performed at a later stage of the sequencing of the human genome. The determination of the allele fre-quency for 60 SNPs resulted in an exclusion of four SNPs (7%) due to not being polymorphic in the Swedish population, despite the thor-

49

ough in silico selection performed. This illustrating that the informa-tion deposited in the databases should be considered with care (Marthet al. 2001; Reich et al. 2003). Analysis of the allele frequency of the SNPs selected in the population they are to be studied in is a good invested time, being able to discard non-informative SNPs at an early stage of the assay.

Applications Pharmacogenetics of antihypertensive response In Study II, we selected a panel of 74 SNPs located in candidate genes of different pathways related to blood pressure regulation. These SNPs were genotyped in 97 hypertensive individuals participating in the SILVHIA trial using the microarray based system with the detection primers immobilised on the surface (Method II). Since multiple SNPs had been selected in several of the genes, we were able to construct haplotypes for eight genes using the expectation maximisation algo-rithm based on the SNP genotypes (Hawley and Kidd 1995). We found that three to six major haplotypes with a frequency > 5% ex-plains the allelic diversity in these genes. By visual inspection of the different haplotypes, we identified two to three haplotype tagging SNPs (htSNPs) (Johnson et al. 2001) for each gene. Analysis of these SNPs would have been enough to capture most of the genetic variation in the genes. The identification and removal of redundant SNPs in haplotype analyses would facilitate the analysis and reduce the cost of the genotyping. This is one of the goals of the HapMap project (HapMapConsortium 2003).

Multiple regression analysis was used to determine the effect of combinations of SNP genotypes on the blood pressure response during antihypertensive treatment with the angiotensin II type 1 receptor antagonist irbesartan, or the 1-adrenergic receptor blocker atenolol in the SILVHIA participants. SNPs with a p-value < 0.10 determined with ANOVA, were entered as independent variables in a forward stepwise multiple regression model with one of the possible combina-tions of systolic or diastolic blood pressure (SBP and DBP respectively) and treatment as dependent variable. This analysis identified combina-tions of four to five SNP genotypes that explain 44-56% of the reduc-tion in SBP and DBP for the two treatment groups (Table 7). As an example, the individuals with the genotype combination for the profile explaining the change in systolic blood pressure upon irbesartan treat-ment showed an average reduction of 45 mmHg, compared to 11 mmHg in individuals with other genotype combinations.

50

Table 7. SNP profiles explaining the blood pressure response to antihypertensive treatment during the SILVHIA trial.

SNP*Genotypea Regression

coefficientb P-value Model R2

SBP Irbesartan

APOA A1449G AA 0.095 <0.001

CYP11B2 T267C TT -0.091 <0.01

EDNRB G40A AA 0.102 <0.01

ENOS G498A GG 0.109 <0.01 0.51

SBP Atenolol

ADRA2A G278T GG -0.18 <0.001

ADRB2 G1342C CC -0.068 <0.01

AGT C1015T CC 0.14 <0.001

EDNRB G40A GG -0.050 <0.01

ENOS G498A AA -0.074 <0.001 0.56

DBP Irbesartan

APOA A1449G AA 0.10 <0.001

ACE A12257G GG -0.073 <0.01

AGT C1198T CC 0.10 <0.01

LIPC A110G GG -0.079 <0.01 0.44

DBP Atenolol

ADRB2 G1309A AA 0.053 <0.01

ENOS A2996G AA 0.094 <0.001

ADRA2A G1817A AA -0.15 <0.01

LIPC A110G GG 0.12 <0.001

EDNRB G40A GG -0.049 <0.01 0.50 a Genotypes with p < 0.10 were entered into the multivariate model and those with p < 0.01 were retained. b A positive regression coefficient should be regarded as a less pronounced reduction on, in some cases, even an increase in the blood pressure compared to those carrying the other allele. Conversely, a negative regression coefficient indicates a more pro-nounced reduction in blood pressure compared to those carrying the other allele.

No SNPs in the receptors of the drugs were identified in the profiles. This might be a result of other steps in the action of the drug being more important, or that the relevant SNP was not included in the panel. To cover the allelic heterogeneity of the genes involved in the action of the drug would require a substantial effort to identify all

51

SNPs present in the genes and an extensive genotyping effort. How-ever, this may be possible in the near future using methods with a ge-nome scale throughput and using htSNPs identified by the HapMap project.

In Study IIIa we used the genotype data from Study II to determine the effect of the genotype for individual SNPs involved in the RAAS in relation to the blood pressure change upon antihypertensive treatment. We found that two SNPs, the AGT T1198C causing a methionine to threonine substitution at amino acid 235 of the protein, and the AGT A1218G, located six nucleotides upstream the transcription start site of the gene, were related to the reduction in SBP during atenolol treatment (Figure 13). The RAAS plays an important role in the blood pressure regulation and many antihypertensive drugs target this sys-tem. A good hypothesis that SNPs in these genes would affect the blood pressure response to antihypertensive treatment, especially for the angiotensin II type 1 receptor blocker irbesartan. The SNP AGT T1198C has been shown to be involved in hypertension by another group (Jeunemaitre et al. 1992). In addition, it has previously been analysed in a set of SILVHIA samples. The same profile was seen for the blood pressure reduction for the different genotypes, but the asso-ciation was not significant (Kurland et al. 2001).

Figure 13. Change in systolic blood pressure related to genotype for two SNPs in the AGT gene upon treatment with atenolol. Data are mean change in BP ± SEM after 12 weeks of treatment.

In study IIIb, genotypes for fifteen SNPs in genes related to lipid me-tabolism from Study II were analysed to identify association to the change in blood pressure upon treatment with irbesartan or atenolol. Four SNPs (APOA-IV A1449G, APOA-V C31455T, APOB G10108A and APOB C711T) with a genotype pattern related to the change in SBP were identified in the irbesartan treatment group and one SNP (LDLR T16730C) was found in association to the change in SBP related to atenolol treatment. After performing a permutation test, the C711T SNP in the apolipoprotein B gene remained associated to the change in SBP upon irbesartan treatment. The SNP T16730C in the low density lipoprotein receptor gene remained associated to the change in SBP upon atenolol treatment (Figure 14).

52

Figure 14. Relationship between SNP genotype and systolic blood pressure change for the SNPs ApoB C711T and LDLR C16730T after correction for multiple testing by a permutation test. The mean change in BP after 12 weeks ± SEM is plotted.

The use of multiplex genotyping methods to evaluate the effect of a SNP on a phenotype results in a problem with multiple comparisons and the risk of findings by chance in the statistical analysis. This risk can be reduced by the use of a lower significance level (p = 0.01) as in Studies II, IIIa and IIIc, with results that need to be confirmed with a lower number of tests. Another approach is to perform a correction for multiple testing either by a permutation test (Study IIIb) or Bonferroni correction. The permutation test allow for a milder correction of the p-values compared to Bonferroni, since the latter leads to an overcorrec-tion due to assuming all tests to be independent thereby increasing the type II error.

The SILVHIA trial was originally designed to determine the efficacy of antihypertensive treatment on the change in LVM in hypertensive individuals. In Study IIIc the relationship between SNP genotypes and LVM upon treatment with irbesartan or atenolol was determined. Analyses for the genotypes of the individual SNPs were performed for SBP and DBP for the two treatment groups to identify the SNPs to be entered into multiple regression analysis. Combinations of SNP geno-types that could explain the reduction in LVM index (LVMI) after 12 weeks of treatment was identified. Of the fifteen SNPs entered into the multiple regression model, a combination of two SNPs, the AGT T1198C and the APOB G1018A, were identified in relation to the reduction in LVMI in the individuals treated with irbesartan. One SNP in the adrenergic alpha-2A-receptor (ADRA2A G1817A) was identi-fied as associated to the reduction in LVMI during the atenolol treat-ment (Table 8).

53

Table 8. SNP profiles explaining the reduction in left ventricular mass to anti-hypertensive treatment during the SILVHIA trial

GenotypeGenotype Regression

coefficientp-value

Irbesartan treatment

AGT T1198C TT 0.098 0.008

ApoB G10108A GG -0.099 0.01

Atenolol treatment

ADRA2A 1817 GG -0.15 0.006

The result for the AGT T1198C SNP confirms the finding of an asso-ciation to reduction in LVMI upon irbesartan treatment by Kurland etal. (Kurland et al. 2002b). This SNP was also identified in Study IIIa as an individual predictor of the reduction in SBP during atenolol treat-ment. The finding of an effect for different treatments on the change in LVMI and SBP respectively may simply imply that different genetic loci are responsible for the change in blood pressure and LVM.

One negative aspect that accounts for Study II and Studies IIIa-c is the limited number of samples in the SILVHIA trial. Although the individuals are clinically well characterised, the results are inconclusive with p-values barely below 0.05. Although the results in Study II and IIIa-c should be considered as hypothesis generating, needing confirma-tion in a larger sample preferably in a prospective pharmacogenetic study, the feasibility of microarray based genotyping for this kind of analysis has been demonstrated. As more is known regarding SNPs and large scale prospective studies will be performed, the complex inter-play of genetic variants and environment in the individual response to a drug may be elucidated and guide a strategy of individualised treat-ment.

54

CONCLUDING REMARKS

With the complete sequence of the human genome available, studying the effect of single nucleotide polymorphisms in disease phenotypes and drug response has become easier. Research in disease genetics has so far mainly focused on genetic variants in coding parts of the genome and their phenotypic effects, but is now moving towards the identifica-tion of genetic variants in regulatory and non-coding regions of the genome. The complexity of the human proteome also is to be deter-mined. High throughput analysis of genetic variation will soon be technically possible at a whole genome scale. Microarray based assays are well suited for genome wide analyses owing to their miniaturised format, and reaction principles allowing for accurate genotyping al-ready exist. Thus, generating large amounts of genotype data will not be difficult. But the evaluation of the data will be problematic; how should all the data be presented in a comprehensive way, and how should the data be statistically evaluated to answer the hypothesis, while recognising the issue of multiple testing in genome wide studies.

Pharmacogenetics will impact the individual as well as the pharma-ceutical industry. With the increasing knowledge of our genome, iden-tification of the genetic causes underlying drug response is simplified. The possibility to determine an individual’s response to a drug prior to administration would increase the efficacy of treatment by avoiding possible side-effects due to an altered metabolism, limit the time to recognise the most adequate treatment, and thus reduce healthcare cost. The identification of pharmacogenetically important markers has so far been limited by small sample sizes resulting in inconclusive re-sults. Well designed prospective studies are needed to identify the most relevant pharmacogenetic markers. In the pharmaceutical indus-try, pharmacokinetic studies are included in the early stages of the drug development process to identify possible side-effects due to ge-netic variation between individuals. Pharmacogenetics offers the prom-ise of a shift from diagnosis of disease to prediction; and from therapy to prevention.

55

ACKNOWLEDGEMENTS

The work of this thesis was performed in the group of Molecular Medicine at the Department of Medical Sciences, Uppsala University.

There are numerous people who have made contributions in one way or the other making this thesis possible. I am grateful to each and every one of you! There are some persons that I especially would like to ac-knowledge;

First of all, my supervisor Ann-Christine Syvänen for all your knowl-edge in various fields and for giving me the opportunity to perform this work in your group.

Present and past friends and colleagues in the group of Molecular Medicine, you have all been great!! Especially I would like to thank Lovisa for being a good friend and colleague always being positive and having time to answer questions, thank you also for providing valuable input on this thesis; Katarina for all the fun we have had when you lived here in Uppsala, and for being a really good friend; Lotta O and Karin J for friendship and sharing office in the past; Marie for being the best team mate in the lab during many minisequencing reactions and for keeping secrets!; my fellow PhD students Mona, Snaevar and An-dreas D for fun and interesting discussions; Lotta R for commenting on this thesis; Kristina, Tomas, Caisa, Annika M, Danne, Mats, Andreas B, Lars B, Anki, Lillebil, Raul, Annika A, Lili, Johanna and Erik for being such great colleagues.

All collaborators in the projects; especially Lars Lind and Lisa Kurland for introducing me to the complexity of hypertension and the field of pharmacogenetics, and Lisa for providing valuable comments on this thesis. Thomas Kahan for giving me access to the SILVHIA trial.

56

Each and everyone at research department II, Uppsala University Hos-pital for providing a cheerful atmosphere to work in, although we are all longing for the new building! Especially I would like to acknowl-edge Mari-Anne, Margareta E and Åsa H for always being helpful and friendly.

Past and present members of the Molecular Medicine group at the Department of Genetics and Pathology headed by Ulf Landegren for interesting seminars and discussions at our joint group meetings.

My friends; Christine for our many Saturdays with shopping and “na-tionsfika”, and for being a really good friend; Petra for keeping me up to date with the skiing in Innsbruck, although I have not tried it yet; Marie-Louise, Linda, Jenny B, Emma, Sofia, Jenny G, Cecilia and Anna-Karin for dinners talking about everything but research! I hope that I will have time to spend much more time with you all, now that I have finished this thesis!!!

This work was financially supported by grants from the Swedish Re-search Council (VR-M & VR-NT) and K&A Wallenbergs Stiftelse (WCN).

Per for always being there for me in good times and bad times!

Sist men inte minst min familj; Mamma, Pappa, Helena och Stefan, Magnus och Jenny, för att ni i varje fall försökt förstå vad jag gjort un-der dessa år. This is it!!!

Uppsala, March 28th 2004

57

REFERENCES

Ahmadian, A., Gharizadeh, B., O'Meara, D., Odeberg, J. and Lundeberg, J. (2001). Genotyping by apyrase-mediated allele-specific extension. Nucleic Acids Res29(24):e121.

Aithal, G. P., Day, C. P., Kesteven, P. J. and Daly, A. K. (1999). Association of poly-morphisms in the cytochrome P450 CYP2C9 with warfarin dose requirement and risk of bleeding complications. Lancet 353(9154):717-9.

Albert, T. J., Norton, J., Ott, M., Richmond, T., Nuwaysir, K., Nuwaysir, E. F., Stengele, K. P. and Green, R. D. (2003). Light-directed 5'-->3' synthesis of com-plex oligonucleotide microarrays. Nucleic Acids Res 31(7):e35.

Alsmadi, O. A., Bornarth, C. J., Song, W., Wisniewski, M., Du, J., Brockman, J. P., Faruqi, A. F., Hosono, S., Sun, Z., Du, Y., et al. (2003). High accuracy genotyping directly from genomic DNA using a rolling circle amplification based assay. BMC Genomics 4(1):21.

Alving, A. S., Carson, P. E., Flanagan, C. L. and Ickes, C. E. (1956). Enzymatic defi-ciency in primaquine-sensitive erythrocytes. Science 124(3220):484-5.

Arava, Y., Wang, Y., Storey, J. D., Liu, C. L., Brown, P. O. and Herschlag, D. (2003). Genome-wide analysis of mRNA translation profiles in Saccharomyces cerevisiae. Proc Natl Acad Sci U S A 100(7):3889-94.

Ardlie, K. G., Kruglyak, L. and Seielstad, M. (2002). Patterns of linkage disequilibrium in the human genome. Nat Rev Genet 3(4):299-309.

Arnheim, N., Strange, C. and Erlich, H. (1985). Use of pooled DNA samples to detect linkage disequilibrium of polymorphic restriction fragments and human disease: studies of the HLA class II loci. Proc Natl Acad Sci U S A 82(20):6970-4.

Avery, O., MacLeod, C. and McCarty, M. (1944). Studies on the chemical nature of the substance inducing transformation in Pneumococcal types. Journal of Exp Med79(137-158.

Banér, J., Isaksson, A., Waldenstrom, E., Jarvius, J., Landegren, U. and Nilsson, M. (2003). Parallel gene analysis with allele-specific padlock probes and tag microar-rays. Nucleic Acids Res 31(17):e103.

Barczak, A., Rodriguez, M. W., Hanspers, K., Koth, L. L., Tai, Y. C., Bolstad, B. M., Speed, T. P. and Erle, D. J. (2003). Spotted long oligonucleotide arrays for human gene expression analysis. Genome Res 13(7):1775-85.

Baum, M., Bielau, S., Rittner, N., Schmid, K., Eggelbusch, K., Dahms, M., Schlauers-bach, A., Tahedl, H., Beier, M., Guimil, R., et al. (2003). Validation of a novel, fully integrated and flexible microarray benchtop facility for gene expression pro-filing. Nucleic Acids Res 31(23):e151.

Beier, M. and Hoheisel, J. D. (2000). Production by quantitative photolithographic synthesis of individually quality checked DNA microarrays. Nucleic Acids Res28(4):e11.

Beier, M. and Hoheisel, J. D. (2002). Analysis of DNA-microarrays produced by in-verse in situ oligonucleotide synthesis. J Biotechnol 94(1):15-22.

58

Bell, P. A., Chaturvedi, S., Gelfand, C. A., Huang, C. Y., Kochersperger, M., Kopla, R., Modica, F., Pohl, M., Varde, S., Zhao, R., et al. (2002). SNPstream UHT: ul-tra-high throughput SNP genotyping for pharmacogenomics and drug discovery. Biotechniques 30(S70-77.

Blanchard, A. P., Kaiser, R. J. and Hood, L. E. (1996). High-density oligonucleotide arrays. Biosensors & Bioelectronics 11(6-7):687-690.

Botstein, D. and Risch, N. (2003). Discovering genotypes underlying human pheno-types: past successes for mendelian disease, future approaches for complex disease. Nat Genet 33(Suppl):228-37.

Brandt, O., Feldner, J., Stephan, A., Schroder, M., Schnolzer, M., Arlinghaus, H. F., Hoheisel, J. D. and Jacob, A. (2003). PNA microarrays for hybridisation of unla-belled DNA samples. Nucleic Acids Res 31(19):e119.

Broude, N. E., Zhang, L., Woodward, K., Englert, D. and Cantor, C. R. (2001). Mul-tiplex allele-specific target amplification based on PCR suppression. Proc Natl Acad Sci U S A 98(1):206-11.

Broude, N. E., Woodward, K., Cavallo, R., Cantor, C. R. and Englert, D. (2001). DNA microarrays with stem-loop DNA probes: preparation and applications. Nu-cleic Acids Res 29(19):e92.

Brownie, J., Shawcross, S., Theaker, J., Whitcombe, D., Ferrie, R., Newton, C. and Little, S. (1997). The elimination of primer-dimer accumulation in PCR. NucleicAcids Res 25(16):3235-41.

Buetow, K. H., Edmonson, M., MacDonald, R., Clifford, R., Yip, P., Kelley, J., Little, D. P., Strausberg, R., Koester, H., Cantor, C. R., et al. (2001). High-throughput development and characterization of a genomewide collection of gene-based single nucleotide polymorphism markers by chip-based matrix-assisted laser desorp-tion/ionization time-of-flight mass spectrometry. Proc Natl Acad Sci U S A98(2):581-4.

Busti, E., Bordoni, R., Castiglioni, B., Monciardini, P., Sosio, M., Donadio, S., Conso-landi, C., Rossi Bernardi, L., Battaglia, C. and De Bellis, G. (2002). Bacterial dis-crimination by means of a universal array approach mediated by LDR (ligase de-tection reaction). BMC Microbiol 2(1):27.

Cai, H., White, P. S., Torney, D., Deshpande, A., Wang, Z., Marrone, B. and Nolan, J. P. (2000). Flow cytometry-based minisequencing: a new platform for high-throughput single-nucleotide polymorphism scoring. Genomics 66(2):135-43.

Callow, M. J., Drmanac, S. and Drmanac, R. (2004). Selective DNA amplification from complex genomes using universal double-sided adapters. Nucleic Acids Res32(2):e21.

Cardon, L. R. and Abecasis, G. R. (2003). Using haplotype blocks to map human complex trait loci. Trends Genet 19(3):135-40.

Chee, M., Yang, R., Hubbell, E., Berno, A., Huang, X. C., Stern, D., Winkler, J., Lockhart, D. J., Morris, M. S. and Fodor, S. P. (1996). Accessing genetic informa-tion with high-density DNA arrays. Science 274(5287):610-4.

Chen, J., Iannone, M. A., Li, M. S., Taylor, J. D., Rivers, P., Nelsen, A. J., Slentz-Kesler, K. A., Roses, A. and Weiner, M. P. (2000). A microsphere-based assay for multiplexed single nucleotide polymorphism analysis using single base chain ex-tension. Genome Res 10(4):549-57.

Churchill, G. A. and Doerge, R. W. (1994). Empirical threshold values for quantita-tive trait mapping. Genetics 138(3):963-71.

Clark, A. G. (2003). Finding genes underlying risk of complex disease by linkage dis-equilibrium mapping. Curr Opin Genet Dev 13(3):296-302.

Corder, E. H., Saunders, A. M., Strittmatter, W. J., Schmechel, D. E., Gaskell, P. C., Small, G. W., Roses, A. D., Haines, J. L. and Pericak-Vance, M. A. (1993). Gene

59

dose of apolipoprotein E type 4 allele and the risk of Alzheimer's disease in late onset families. Science 261(5123):921-3.

Cotton, R. G., Rodrigues, N. R. and Campbell, R. D. (1988). Reactivity of cytosine and thymine in single-base-pair mismatches with hydroxylamine and osmium tetroxide and its application to the study of mutations. Proc Natl Acad Sci U S A85(12):4397-401.

Croucher, P. J., Mascheretti, S., Hampe, J., Huse, K., Frenzel, H., Stoll, M., Lu, T., Nikolaus, S., Yang, S. K., Krawczak, M., et al. (2003). Haplotype structure and as-sociation to Crohn's disease of CARD15 mutations in two ethnically divergent populations. Eur J Hum Genet 11(1):6-16.

Dahlof, B., Devereux, R. B., Kjeldsen, S. E., Julius, S., Beevers, G., Faire, U., Fy-hrquist, F., Ibsen, H., Kristiansson, K., Lederballe-Pedersen, O., et al. (2002). Cardiovascular morbidity and mortality in the Losartan Intervention For Endpoint reduction in hypertension study (LIFE): a randomised trial against atenolol. Lancet359(9311):995-1003.

Daly, M. J., Rioux, J. D., Schaffner, S. F., Hudson, T. J. and Lander, E. S. (2001). High-resolution haplotype structure in the human genome. Nat Genet 29(2):229-32.

Day, P. J., Flora, P. S., Fox, J. E. and Walker, M. R. (1991). Immobilization of polynu-cleotides on magnetic particles. Factors influencing hybridization efficiency. Bio-chem J 278(735-40.

Deloukas, P. and Bentley, D. (2004). The HapMap project and its application to ge-netic studies of drug response. Pharmacogenomics J 4(2):88-90.

Delrio-Lafreniere, S. A., Browning, M. K. and McGlennen, R. C. (2004). Low-density addressable array for the detection and typing of the human papillomavirus. Diagn Microbiol Infect Dis 48(1):23-31.

Devereux, R. B., de Simone, G., Koren, M. J., Roman, M. J. and Laragh, J. H. (1991). Left ventricular mass as a predictor of development of hypertension. Am J Hyper-tens 4(11):603S-607S.

Don, R. H., Cox, P. T., Wainwright, B. J., Baker, K. and Mattick, J. S. (1991). 'Touchdown' PCR to circumvent spurious priming during gene amplification. Nu-cleic Acids Res 19(14):4008.

Drmanac, R., Labat, I., Brukner, I. and Crkvenjakov, R. (1989). Sequencing of megabase plus DNA by hybridization: theory of the method. Genomics 4(2):114-28.

Drmanac, R., Drmanac, S., Strezoska, Z., Paunesku, T., Labat, I., Zeremski, M., Snoddy, J., Funkhouser, W. K., Koop, B., Hood, L., et al. (1993). DNA sequence determination by hybridization: a strategy for efficient large-scale sequencing. Sci-ence 260(5114):1649-52.

Edman, C. F., Raymond, D. E., Wu, D. J., Tu, E., Sosnowski, R. G., Butler, W. F., Nerenberg, M. and Heller, M. J. (1997). Electric field directed nucleic acid hy-bridization on microchips. Nucleic Acids Res 25(24):4907-14.

Eickhoff, H., Konthur, Z., Lueking, A., Lehrach, H., Walter, G., Nordhoff, E., Nyar-sik, L. and Bussow, K. (2002). Protein array technology: the tool to bridge genom-ics and proteomics. Adv Biochem Eng Biotechnol 77(103-12.

Evans, W. E. and Relling, M. V. (1999). Pharmacogenomics: translating functional genomics into rational therapeutics. Science 286(5439):487-91.

Evans, W. E. (2003). Pharmacogenomics: marshalling the human genome to individu-alise drug therapy. Gut 52(Suppl 2):ii10-8.

Fagard, R. H., Loos, R. J., Beunen, G., Derom, C. and Vlietinck, R. (2003). Influence of chorionicity on the heritability estimates of blood pressure: a study in twins. JHypertens 21(7):1313-8.

60

Fan, J. B., Chen, X., Halushka, M. K., Berno, A., Huang, X., Ryder, T., Lipshutz, R. J., Lockhart, D. J. and Chakravarti, A. (2000). Parallel genotyping of human SNPs using generic high-density oligonucleotide tag arrays. Genome Res 10(6):853-60.

Fodor, S. P., Read, J. L., Pirrung, M. C., Stryer, L., Lu, A. T. and Solas, D. (1991). Light-directed, spatially addressable parallel chemical synthesis. Science251(4995):767-73.

Fortina, P., Delgrosso, K., Sakazume, T., Santacroce, R., Moutereau, S., Su, H. J., Graves, D., McKenzie, S. and Surrey, S. (2000). Simple two-color array-based ap-proach for mutation detection. Eur J Hum Genet 8(11):884-94.

Fotin, A. V., Drobyshev, A. L., Proudnikov, D. Y., Perov, A. N. and Mirzabekov, A. D. (1998). Parallel thermodynamic analysis of duplexes on oligodeoxyribonucleo-tide microchips. Nucleic Acids Res 26(6):1515-21.

Gabriel, S. B., Schaffner, S. F., Nguyen, H., Moore, J. M., Roy, J., Blumenstiel, B., Higgins, J., DeFelice, M., Lochner, A., Faggart, M., et al. (2002). The structure of haplotype blocks in the human genome. Science 296(5576):2225-9.

Gerry, N. P., Witowski, N. E., Day, J., Hammer, R. P., Barany, G. and Barany, F. (1999). Universal DNA microarray method for multiplex detection of low abun-dance point mutations. J Mol Biol 292(2):251-62.

Giallourakis, C., Stoll, M., Miller, K., Hampe, J., Lander, E. S., Daly, M. J., Schreiber, S. and Rioux, J. D. (2003). IBD5 is a general risk factor for inflammatory bowel disease: replication of association with Crohn disease and identification of a novel association with ulcerative colitis. Am J Hum Genet 73(1):205-11.

Gingeras, T. R., Kwoh, D. Y. and Davis, G. R. (1987). Hybridization properties of immobilized nucleic acids. Nucleic Acids Res 15(13):5373-90.

Goldstein, D. B. (2001). Islands of linkage disequilibrium. Nat Genet 29(2):109-11. Grant, D. M., Goodfellow, G. H., Sugamori, K. and Durette, K. (2000). Pharmacoge-

netics of the human arylamine N-acetyltransferases. Pharmacology 61(3):204-11. Griffin, T. J., Tang, W. and Smith, L. M. (1997). Genetic analysis by peptide nucleic

acid affinity MALDI-TOF mass spectrometry. Nat Biotechnol 15(13):1368-72. Gunderson, K. L., Huang, X. C., Morris, M. S., Lipshutz, R. J., Lockhart, D. J. and

Chee, M. S. (1998). Mutation detection by ligation to complete n-mer DNA ar-rays. Genome Res 8(11):1142-53.

Guo, Z., Guilfoyle, R. A., Thiel, A. J., Wang, R. and Smith, L. M. (1994). Direct fluorescence analysis of genetic polymorphisms by hybridization with oligonucleo-tide arrays on glass supports. Nucleic Acids Res 22(24):5456-65.

Guschin, D., Yershov, G., Zaslavsky, A., Gemmell, A., Shick, V., Proudnikov, D., Arenkov, P. and Mirzabekov, A. (1997). Manual manufacturing of oligonucleo-tide, DNA, and protein microchips. Anal Biochem 250(2):203-11.

Guyton, A. C., Hall, J. E., Lohmeier, T. E., Jackson, T. E. and Kastner, P. R. (1981). Blood pressure regulation: basic concepts. Fed Proc 40(8):2252-6.

Hacia, J. G., Brody, L. C., Chee, M. S., Fodor, S. P. and Collins, F. S. (1996). Detec-tion of heterozygous mutations in BRCA1 using high density oligonucleotide ar-rays and two-colour fluorescence analysis. Nat Genet 14(4):441-7.

Hacia, J. G., Sun, B., Hunt, N., Edgemon, K., Mosbrook, D., Robbins, C., Fodor, S. P., Tagle, D. A. and Collins, F. S. (1998). Strategies for mutational analysis of the large multiexon ATM gene using high-density oligonucleotide arrays. Genome Res8(12):1245-58.

HapMapConsortium. (2003). The International HapMap Project. Nature426(6968):789-96.

Hardenbol, P., Baner, J., Jain, M., Nilsson, M., Namsaraev, E. A., Karlin-Neumann, G. A., Fakhrai-Rad, H., Ronaghi, M., Willis, T. D., Landegren, U., et al. (2003). Mul-

61

tiplexed genotyping with sequence-tagged molecular inversion probes. Nat Bio-technol 21(6):673-8.

Hawley, M. E. and Kidd, K. K. (1995). HAPLO: a program using the EM algorithm to estimate the frequencies of multi-site haplotypes. J Hered 86(5):409-11.

Head, S. R., Rogers, Y. H., Parikh, K., Lan, G., Anderson, S., Goelet, P. and Boyce-Jacino, M. T. (1997). Nested genetic bit analysis (N-GBA) for mutation detection in the p53 tumor suppressor gene. Nucleic Acids Res 25(24):5065-71.

Helgadottir, A., Manolescu, A., Thorleifsson, G., Gretarsdottir, S., Jonsdottir, H., Thorsteinsdottir, U., Samani, N. J., Gudmundsson, G., Grant, S. F., Thorgeirsson, G., et al. (2004). The gene encoding 5-lipoxygenase activating protein confers risk of myocardial infarction and stroke. Nat Genet 36(3):233-9.

Hirschhorn, J. N., Sklar, P., Lindblad-Toh, K., Lim, Y. M., Ruiz-Gutierrez, M., Bolk, S., Langhorst, B., Schaffner, S., Winchester, E. and Lander, E. S. (2000). SBE-TAGS: an array-based method for efficient single-nucleotide polymorphism geno-typing. Proc Natl Acad Sci U S A 97(22):12164-9.

Hochberg, E. P., Miklos, D. B., Neuberg, D., Eichner, D. A., McLaughlin, S. F., Mattes-Ritz, A., Alyea, E. P., Antin, J. H., Soiffer, R. J. and Ritz, J. (2003). A novel rapid single nucleotide polymorphism (SNP)-based method for assessment of hematopoietic chimerism after allogeneic stem cell transplantation. Blood101(1):363-9.

Holland, P. M., Abramson, R. D., Watson, R. and Gelfand, D. H. (1991). Detection of specific polymerase chain reaction product by utilizing the 5'----3' exonuclease ac-tivity of Thermus aquaticus DNA polymerase. Proc Natl Acad Sci U S A88(16):7276-80.

Horikawa, Y., Oda, N., Cox, N. J., Li, X., Orho-Melander, M., Hara, M., Hinokio, Y., Lindner, T. H., Mashima, H., Schwarz, P. E., et al. (2000). Genetic variation in the gene encoding calpain-10 is associated with type 2 diabetes mellitus. NatGenet 26(2):163-75.

Hudson, T. J. (2003). Wanted: regulatory SNPs. Nat Genet 33(4):439-40. Hughes, H. B., Biehl, J. P., Jones, A. P. and Schmidt, L. H. (1954). Metabolism of

isoniazid in man as related to the occurrence of peripheral neuritis. Am Rev Tuberc70(2):266-73.

Hughes, T. R., Mao, M., Jones, A. R., Burchard, J., Marton, M. J., Shannon, K. W., Lefkowitz, S. M., Ziman, M., Schelter, J. M., Meyer, M. R., et al. (2001). Expres-sion profiling using microarrays fabricated by an ink-jet oligonucleotide synthe-sizer. Nat Biotechnol 19(4):342-7.

Hugot, J. P., Chamaillard, M., Zouali, H., Lesage, S., Cezard, J. P., Belaiche, J., Almer, S., Tysk, C., O'Morain, C. A., Gassull, M., et al. (2001). Association of NOD2 leucine-rich repeat variants with susceptibility to Crohn's disease. Nature411(6837):599-603.

Huntington'sDiseaseCollaborativeGroup. (1993). A novel gene containing a trinucleo-tide repeat that is expanded and unstable on Huntington's disease chromosomes. The Huntington's Disease Collaborative Research Group. Cell 72(6):971-83.

Iannone, M. A., Taylor, J. D., Chen, J., Li, M. S., Rivers, P., Slentz-Kesler, K. A. and Weiner, M. P. (2000). Multiplexed single nucleotide polymorphism genotyping by oligonucleotide ligation and flow cytometry. Cytometry 39(2):131-40.

Ingelman-Sundberg, M. (2004). Human drug metabolising cytochrome P450 enzymes: properties and polymorphisms. Naunyn Schmiedebergs Arch Pharmacol 369(1):89-104.

Jeffreys, A. J., Kauppi, L. and Neumann, R. (2001). Intensely punctate meiotic re-combination in the class II region of the major histocompatibility complex. NatGenet 29(2):217-22.

62

Jeunemaitre, X., Soubrier, F., Kotelevtsev, Y. V., Lifton, R. P., Williams, C. S., Charru, A., Hunt, S. C., Hopkins, P. N., Williams, R. R., Lalouel, J. M., et al.(1992). Molecular basis of human hypertension: role of angiotensinogen. Cell71(1):169-80.

Jobs, M., Howell, W. M., Stromqvist, L., Mayr, T. and Brookes, A. J. (2003). DASH-2: flexible, low-cost, and high-throughput SNP genotyping by dynamic allele-specific hybridization on membrane arrays. Genome Res 13(5):916-24.

Johnson, G. C., Esposito, L., Barratt, B. J., Smith, A. N., Heward, J., Di Genova, G., Ueda, H., Cordell, H. J., Eaves, I. A., Dudbridge, F., et al. (2001). Haplotype tag-ging for the identification of common disease genes. Nat Genet 29(2):233-7.

Johnson, J. A. and Lima, J. J. (2003). Drug receptor/effector polymorphisms and pharmacogenetics: current status and challenges. Pharmacogenetics 13(9):525-34.

Johnson, J. A. (2003). Pharmacogenetics: potential for individualized drug therapy through genetics. Trends Genet 19(11):660-6.

Joos, B., Kuster, H. and Cone, R. (1997). Covalent attachment of hybridizable oli-gonucleotides to glass supports. Anal Biochem 247(1):96-101.

Kaderali, L., Deshpande, A., Nolan, J. P. and White, P. S. (2003). Primer-design for multiplexed genotyping. Nucleic Acids Res 31(6):1796-802.

Kajiyama, T., Miyahara, Y., Kricka, L. J., Wilding, P., Graves, D. J., Surrey, S. and Fortina, P. (2003). Genotyping on a thermal gradient DNA chip. Genome Res13(3):467-75.

Karttunen, L., Lonnqvist, L., Godfrey, M., Peltonen, L. and Syvanen, A. C. (1996). An accurate method for comparing transcript levels of two alleles or highly ho-mologous genes: application to fibrillin transcripts in Marfan patients' fibroblasts. Genome Res 6(5):392-403.

Katsanis, N., Worley, K. C. and Lupski, J. R. (2001). An evaluation of the draft hu-man genome sequence. Nat Genet 29(1):88-91.

Ke, X., Hunt, S., Tapper, W., Lawrence, R., Stavrides, G., Ghori, J., Whittaker, P., Collins, A., Morris, A. P., Bentley, D., et al. (2004). The impact of SNP density on fine-scale patterns of linkage disequilibrium. Hum Mol Genet 13(6):577-88.

Kennedy, G. C., Matsuzaki, H., Dong, S., Liu, W. M., Huang, J., Liu, G., Su, X., Cao, M., Chen, W., Zhang, J., et al. (2003). Large-scale genotyping of complex DNA. Nat Biotechnol 21(10):1233-7.

Kerem, B. S., Buchanan, J. A., Durie, P., Corey, M. L., Levison, H., Rommens, J. M., Buchwald, M. and Tsui, L. C. (1989). DNA marker haplotype association with pancreatic sufficiency in cystic fibrosis. Am J Hum Genet 44(6):827-34.

Klein, T. E. and Altman, R. B. (2004). PharmGKB: the pharmacogenetics and phar-macogenomics knowledge base. Pharmacogenomics J 4(1):1.

Koopmans, R. P., Insel, P. A. and Michel, M. C. (2003). Pharmacogenetics of hyper-tension treatment: a structured review. Pharmacogenetics 13(12):705-713.

Koren, M. J., Devereux, R. B., Casale, P. N., Savage, D. D. and Laragh, J. H. (1991). Relation of left ventricular mass and geometry to morbidity and mortality in un-complicated essential hypertension. Ann Intern Med 114(5):345-52.

Koren, M. J., Ulin, R. J., Koren, A. T., Laragh, J. H. and Devereux, R. B. (2002). Left ventricular mass change during treatment and outcome in patients with essential hypertension. Am J Hypertens 15(12):1021-8.

Kurg, A., Tonisson, N., Georgiou, I., Shumaker, J., Tollett, J. and Metspalu, A. (2000). Arrayed primer extension: solid-phase four-color DNA resequencing and mutation detection technology. Genet Test 4(1):1-7.

Kurland, L., Melhus, H., Karlsson, J., Kahan, T., Malmqvist, K., Ohman, K. P., Nyst-rom, F., Hagg, A. and Lind, L. (2001). Angiotensin converting enzyme gene

63

polymorphism predicts blood pressure response to angiotensin II receptor type 1 antagonist treatment in hypertensive patients. J Hypertens 19(10):1783-7.

Kurland, L., Melhus, H., Karlsson, J., Kahan, T., Malmqvist, K., Ohman, P., Nystrom, F., Hagg, A. and Lind, L. (2002a). Aldosterone synthase (CYP11B2) -344 C/T polymorphism is related to antihypertensive response: result from the Swedish Ir-besartan Left Ventricular Hypertrophy Investigation versus Atenolol (SILVHIA) trial. Am J Hypertens 15(5):389-93.

Kurland, L., Melhus, H., Karlsson, J., Kahan, T., Malmqvist, K., Ohman, P., Nystrom, F., Hagg, A. and Lind, L. (2002b). Polymorphisms in the angiotensinogen and an-giotensin II type 1 receptor gene are related to change in left ventricular mass dur-ing antihypertensive treatment: results from the Swedish Irbesartan Left Ventricu-lar Hypertrophy Investigation versus Atenolol (SILVHIA) trial. J Hypertens20(4):657-63.

Kurland, L., Liljedahl, U., Karlsson, J., Kahan, T., Malmqvist, K., Melhus, H., Syva-nen, A. C. and Lind, L. (2004). Angiotensinogen gene polymorphisms: relation-ship to blood pressure response to antihypertensive treatment. Results from the Swedish Irbesartan Left Ventricular Hypertrophy Investigation vs Atenolol (SILVHIA) trial. Am J Hypertens 17(1):8-13.

Kwiatkowski, M., Samiotaki, M., Lamminmaki, U., Mukkala, V. M. and Landegren, U. (1994). Solid-phase synthesis of chelate-labelled oligonucleotides: application in triple-color ligase-mediated gene analysis. Nucleic Acids Res 22(13):2604-11.

Kwiatkowski, M., Fredriksson, S., Isaksson, A., Nilsson, M. and Landegren, U. (1999). Inversion of in situ synthesized oligonucleotides: improved reagents for hybridiza-tion and primer extension in DNA microarrays. Nucleic Acids Res 27(24):4710-4.

Laan, M., Gron-Virta, K., Salo, A., Aula, P., Peltonen, L., Palotie, A. and Syvanen, A. C. (1995). Solid-phase minisequencing confirmed by FISH analysis in determina-tion of gene copy number. Hum Genet 96(3):275-80.

Lagerstrom-Fermer, M., Olsson, C., Forsgren, L. and Syvanen, A. C. (2001). Hetero-plasmy of the human mtDNA control region remains constant during life. Am J Hum Genet 68(5):1299-301.

Landegren, U., Kaiser, R., Sanders, J. and Hood, L. (1988). A ligase-mediated gene detection technique. Science 241(4869):1077-80.

Lander, E. S., Linton, L. M., Birren, B., Nusbaum, C., Zody, M. C., Baldwin, J., Devon, K., Dewar, K., Doyle, M., FitzHugh, W., et al. (2001). Initial sequencing and analysis of the human genome. Nature 409(6822):860-921.

Landi, S., Gemignani, F., Gioia-Patricola, L., Chabrier, A. and Canzian, F. (2003). Evaluation of a microarray for genotyping polymorphisms related to xenobiotic metabolism and DNA repair. Biotechniques 35(4):816-20, 822, 824-7.

Latorra, D., Campbell, K., Wolter, A. and Hurley, J. M. (2003). Enhanced allele-specific PCR discrimination in SNP genotyping using 3' locked nucleic acid (LNA) primers. Hum Mutat 22(1):79-85.

Lazarou, J., Pomeranz, B. H. and Corey, P. N. (1998). Incidence of adverse drug reac-tions in hospitalized patients: a meta-analysis of prospective studies. Jama279(15):1200-5.

Lehmann, H. and Ryan, E. (1956). The familial incidence of low pseudocholinesterase level. Lancet 271(6934):124.

Levine, M. and Tjian, R. (2003). Transcription regulation and animal diversity. Nature424(6945):147-51.

Levy, D., Garrison, R. J., Savage, D. D., Kannel, W. B. and Castelli, W. P. (1990). Prognostic implications of echocardiographically determined left ventricular mass in the Framingham Heart Study. N Engl J Med 322(22):1561-6.

64

Lifton, R. P., Gharavi, A. G. and Geller, D. S. (2001). Molecular mechanisms of hu-man hypertension. Cell 104(4):545-56.

Liggett, S. B. (2000). The pharmacogenetics of beta2-adrenergic receptors: relevance to asthma. J Allergy Clin Immunol 105(2 Pt 2):S487-92.

Lindroos, K., Liljedahl, U., Raitio, M. and Syvanen, A. C. (2001). Minisequencing on oligonucleotide microarrays: comparison of immobilisation chemistries. Nucleic Acids Res 29(13):e69.

Lindroos, K., Sigurdsson, S., Johansson, K., Ronnblom, L. and Syvanen, A. C. (2002). Multiplex SNP genotyping in pooled DNA samples by a four-colour microarray system. Nucleic Acids Res 30(14):e70.

Livak, K. J., Flood, S. J., Marmaro, J., Giusti, W. and Deetz, K. (1995). Oligonucleo-tides with fluorescent dyes at opposite ends provide a quenched probe system use-ful for detecting PCR product and nucleic acid hybridization. PCR Methods Appl4(6):357-62.

Lizardi, P. M., Huang, X., Zhu, Z., Bray-Ward, P., Thomas, D. C. and Ward, D. C. (1998). Mutation detection and single-molecule counting using isothermal rolling-circle amplification. Nat Genet 19(3):225-32.

Lo, H. S., Wang, Z., Hu, Y., Yang, H. H., Gere, S., Buetow, K. H. and Lee, M. P. (2003). Allelic variation in gene expression is common in the human genome. Ge-nome Res 13(8):1855-62.

Lovmar, L., Fredriksson, M., Liljedahl, U., Sigurdsson, S. and Syvanen, A. C. (2003). Quantitative evaluation by minisequencing and microarrays reveals accurate mul-tiplexed SNP genotyping of whole genome amplified DNA. Nucleic Acids Res31(21):e129.

Malmqvist, K., Kahan, T., Edner, M., Held, C., Hagg, A., Lind, L., Muller-Brunotte, R., Nystrom, F., Ohman, K. P., Osbakken, M. D., et al. (2001). Regression of left ventricular hypertrophy in human hypertension with irbesartan. J Hypertens19(6):1167-76.

Maniatis, T. and Tasic, B. (2002). Alternative pre-mRNA splicing and proteome ex-pansion in metazoans. Nature 418(6894):236-43.

Marth, G., Yeh, R., Minton, M., Donaldson, R., Li, Q., Duan, S., Davenport, R., Miller, R. D. and Kwok, P. Y. (2001). Single-nucleotide polymorphisms in the public domain: how useful are they? Nat Genet 27(4):371-2.

Marth, G., Schuler, G., Yeh, R., Davenport, R., Agarwala, R., Church, D., Wheelan, S., Baker, J., Ward, M., Kholodov, M., et al. (2003). Sequence variations in the public human genome data reflect a bottlenecked population history. Proc Natl Acad Sci U S A 100(1):376-81.

Matson, R. S., Rampal, J. B. and Coassin, P. J. (1994). Biopolymer synthesis on poly-propylene supports. I. Oligonucleotides. Anal Biochem 217(2):306-10.

Matson, R. S., Rampal, J., Pentoney, S. L., Jr., Anderson, P. D. and Coassin, P. (1995). Biopolymer synthesis on polypropylene supports: oligonucleotide arrays. Anal Bio-chem 224(1):110-6.

Matsuzaki, H., Loi, H., Dong, S., Tsai, Y. Y., Fang, J., Law, J., Di, X., Liu, W. M., Yang, G., Liu, G., et al. (2004). Parallel Genotyping of Over 10,000 SNPs Using a One-Primer Assay on a High-Density Oligonucleotide Array. Genome Res14(3):414-25.

Maxam, A. M. and Gilbert, W. (1977). A new method for sequencing DNA. Proc Natl Acad Sci U S A 74(2):560-4.

McGall, G. H., Barone, A. D., Diggelmann, M., Fodor, S. P. A., Gentalen, E. and Ngo, N. (1997). The efficiency of light-directed synthesis of DNA arrays on glass sub-strates. Journal of the American Chemical Society 119(22):5081-5090.

65

McLeod, H. L., Krynetski, E. Y., Relling, M. V. and Evans, W. E. (2000). Genetic polymorphism of thiopurine methyltransferase and its clinical relevance for child-hood acute lymphoblastic leukemia. Leukemia 14(4):567-72.

Mein, C. A., Caulfield, M. J., Dobson, R. J. and Munroe, P. B. (2004). Genetics of essential hypertension. Hum Mol Genet 13 Suppl 1(R169-75.

Miller, S. A., Dykes, D. D. and Polesky, H. F. (1988). A simple salting out procedure for extracting DNA from human nucleated cells. Nucleic Acids Res 16(3):1215.

Mitra, R. D. and Church, G. M. (1999). In situ localized amplification and contact replication of many individual DNA molecules. Nucleic Acids Res 27(24):e34.

Mitra, R. D., Butty, V. L., Shendure, J., Williams, B. R., Housman, D. E. and Church, G. M. (2003). Digital genotyping and haplotyping with polymerase colonies. Proc Natl Acad Sci U S A 100(10):5926-31.

Modrek, B. and Lee, C. J. (2003). Alternative splicing in the human, mouse and rat genomes is associated with an increased frequency of exon creation and/or loss. Nat Genet 34(2):177-80.

Mohlke, K. L., Erdos, M. R., Scott, L. J., Fingerlin, T. E., Jackson, A. U., Silander, K., Hollstein, P., Boehnke, M. and Collins, F. S. (2002). High-throughput screening for evidence of association by using mass spectrometry genotyping on DNA pools. Proc Natl Acad Sci U S A 99(26):16928-33.

Mullis, K. B. and Faloona, F. A. (1987). Specific synthesis of DNA in vitro via a poly-merase-catalyzed chain reaction. Methods Enzymol 155(335-50.

Murphy, G. M., Jr., Kremer, C., Rodrigues, H. E. and Schatzberg, A. F. (2003). Phar-macogenetics of antidepressant medication intolerance. Am J Psychiatry160(10):1830-5.

Nguyen, H. K. and Southern, E. M. (2000). Minimising the secondary structure of DNA targets by incorporation of a modified deoxynucleoside: implications for nu-cleic acid analysis by hybridisation. Nucleic Acids Res 28(20):3904-9.

Nikiforov, T. T. and Rogers, Y. H. (1995). The use of 96-well polystyrene plates for DNA hybridization-based assays: an evaluation of different approaches to oligonu-cleotide immobilization. Anal Biochem 227(1):201-9.

Nilsson, M., Malmgren, H., Samiotaki, M., Kwiatkowski, M., Chowdhary, B. P. and Landegren, U. (1994). Padlock probes: circularizing oligonucleotides for localized DNA detection. Science 265(5181):2085-8.

Norton, N., Williams, N. M., Williams, H. J., Spurlock, G., Kirov, G., Morris, D. W., Hoogendoorn, B., Owen, M. J. and O'Donovan, M. C. (2002). Universal, robust, highly quantitative SNP allele frequency measurement in DNA pools. Hum Genet110(5):471-8.

Nuwaysir, E. F., Huang, W., Albert, T. J., Singh, J., Nuwaysir, K., Pitas, A., Rich-mond, T., Gorski, T., Berg, J. P., Ballin, J., et al. (2002). Gene expression analysis using oligonucleotide arrays produced by maskless photolithography. Genome Res12(11):1749-55.

Ogura, Y., Bonen, D. K., Inohara, N., Nicolae, D. L., Chen, F. F., Ramos, R., Britton, H., Moran, T., Karaliuskas, R., Duerr, R. H., et al. (2001). A frameshift mutation in NOD2 associated with susceptibility to Crohn's disease. Nature411(6837):603-6.

Oliphant, A., Barker, D. L., Stuelpnagel, J. R. and Chee, M. S. (2002). BeadArray technology: enabling an accurate, cost-effective approach to high-throughput genotyping. Biotechniques Suppl(56-61.

Olsson, C., Waldenstrom, E., Westermark, K., Landegren, U. and Syvanen, A. C. (2000). Determination of the frequencies of ten allelic variants of the Wilson dis-ease gene (ATP7B), in pooled DNA samples. Eur J Hum Genet 8(12):933-8.

66

O'Meara, D., Ahmadian, A., Odeberg, J. and Lundeberg, J. (2002). SNP typing by apyrase-mediated allele-specific primer extension on DNA microarrays. NucleicAcids Res 30(15):e75.

Orum, H., Jakobsen, M. H., Koch, T., Vuust, J. and Borre, M. B. (1999). Detection of the factor V Leiden mutation by direct allele-specific hybridization of PCR ampli-cons to photoimmobilized locked nucleic acids. Clin Chem 45(11):1898-905.

Ozaki, K., Ohnishi, Y., Iida, A., Sekine, A., Yamada, R., Tsunoda, T., Sato, H., Hori, M., Nakamura, Y. and Tanaka, T. (2002). Functional SNPs in the lymphotoxin-alpha gene that are associated with susceptibility to myocardial infarction. NatGenet 32(4):650-4.

Pastinen, T., Kurg, A., Metspalu, A., Peltonen, L. and Syvanen, A. C. (1997). Minise-quencing: a specific tool for DNA analysis and diagnostics on oligonucleotide ar-rays. Genome Res 7(6):606-14.

Pastinen, T., Raitio, M., Lindroos, K., Tainola, P., Peltonen, L. and Syvanen, A. C. (2000). A system for specific, high-throughput genotyping by allele-specific primer extension on microarrays. Genome Res 10(7):1031-42.

Pastinen, T., Perola, M., Ignatius, J., Sabatti, C., Tainola, P., Levander, M., Syvanen, A. C. and Peltonen, L. (2001). Dissecting a population genome for targeted screening of disease mutations. Hum Mol Genet 10(26):2961-72.

Pastinen, T., Sladek, R., Gurd, S., Sammak, A., Ge, B., Lepage, P., Lavergne, K., Vil-leneuve, A., Gaudin, T., Brandstrom, H., et al. (2004). A survey of genetic and epigenetic variation affecting human gene expression. Physiol Genomics 16(2):184-93.

Patil, N., Berno, A. J., Hinds, D. A., Barrett, W. A., Doshi, J. M., Hacker, C. R., Kautzer, C. R., Lee, D. H., Marjoribanks, C., McDonough, D. P., et al. (2001). Blocks of limited haplotype diversity revealed by high-resolution scanning of hu-man chromosome 21. Science 294(5547):1719-23.

Pausova, Z., Gossard, F., Gaudet, D., Tremblay, J., Kotchen, T. A., Cowley, A. W. and Hamet, P. (2001). Heritability estimates of obesity measures in siblings with and without hypertension. Hypertension 38(1):41-7.

Pease, A. C., Solas, D., Sullivan, E. J., Cronin, M. T., Holmes, C. P. and Fodor, S. P. (1994). Light-generated oligonucleotide arrays for rapid DNA sequence analysis. Proc Natl Acad Sci U S A 91(11):5022-6.

Peltonen, L. and McKusick, V. A. (2001). Genomics and medicine. Dissecting human disease in the postgenomic era. Science 291(5507):1224-9.

Phillips, M. S., Lawrence, R., Sachidanandam, R., Morris, A. P., Balding, D. J., Donaldson, M. A., Studebaker, J. F., Ankener, W. M., Alfisi, S. V., Kuo, F. S., et al. (2003). Chromosome-wide distribution of haplotype blocks and the role of re-combination hot spots. Nat Genet 33(3):382-7.

Pritchard, J. K. and Cox, N. J. (2002). The allelic architecture of human disease genes: common disease-common variant. or not? Hum Mol Genet 11(20):2417-23.

Ramakrishnan, R., Dorris, D., Lublinsky, A., Nguyen, A., Domanus, M., Prokhorova, A., Gieser, L., Touma, E., Lockner, R., Tata, M., et al. (2002). An assessment of Motorola CodeLink(TM) microarray performance for gene expression profiling applications. Nucleic Acids Res 30(7):e30.

Rao, K. V., Stevens, P. W., Hall, J. G., Lyamichev, V., Neri, B. P. and Kelso, D. M. (2003). Genotyping single nucleotide polymorphisms directly from genomic DNA by invasive cleavage reaction on microspheres. Nucleic Acids Res 31(11):e66.

Reese, M. O., van Dam, R. M., Scherer, A. and Quake, S. R. (2003). Microfabricated fountain pens for high-density DNA arrays. Genome Res 13(10):2348-52.

67

Reich, D. E., Cargill, M., Bolk, S., Ireland, J., Sabeti, P. C., Richter, D. J., Lavery, T., Kouyoumjian, R., Farhadian, S. F., Ward, R., et al. (2001). Linkage disequilibrium in the human genome. Nature 411(6834):199-204.

Reich, D. E. and Lander, E. S. (2001). On the allelic spectrum of human disease. Trends Genet 17(9):502-10.

Reich, D. E., Gabriel, S. B. and Altshuler, D. (2003). Quality and completeness of SNP databases. Nat Genet 33(4):457-8.

Rioux, J. D., Daly, M. J., Silverberg, M. S., Lindblad, K., Steinhart, H., Cohen, Z., Delmonte, T., Kocher, K., Miller, K., Guschwan, S., et al. (2001). Genetic varia-tion in the 5q31 cytokine gene cluster confers susceptibility to Crohn disease. NatGenet 29(2):223-8.

Roberts, G. C. and Smith, C. W. (2002). Alternative splicing: combinatorial output from the genome. Curr Opin Chem Biol 6(3):375-83.

Rogers, Y. H., Jiang-Baucom, P., Huang, Z. J., Bogdanov, V., Anderson, S. and Boyce-Jacino, M. T. (1999). Immobilization of oligonucleotides onto a glass support via disulfide bonds: A method for preparation of DNA microarrays. Anal Biochem266(1):23-30.

Ross, P., Hall, L., Smirnov, I. and Haff, L. (1998). High level multiplex genotyping by MALDI-TOF mass spectrometry. Nat Biotechnol 16(13):1347-51.

Ross, P. L., Lee, K. and Belgrader, P. (1997). Discrimination of single-nucleotide polymorphisms in human DNA using peptide nucleic acid probes detected by MALDI-TOF mass spectrometry. Anal Chem 69(20):4197-202.

Rozen, S. and Skaletsky, H. (2000). Primer3 on the WWW for general users and for biologist programmers. Methods Mol Biol 132(365-86.

Sachidanandam, R., Weissman, D., Schmidt, S. C., Kakol, J. M., Stein, L. D., Marth, G., Sherry, S., Mullikin, J. C., Mortimore, B. J., Willey, D. L., et al. (2001). A map of human genome sequence variation containing 1.42 million single nucleo-tide polymorphisms. Nature 409(6822):928-33.

Saiki, R. K., Bugawan, T. L., Horn, G. T., Mullis, K. B. and Erlich, H. A. (1986). Analysis of enzymatically amplified beta-globin and HLA-DQ alpha DNA with al-lele-specific oligonucleotide probes. Nature 324(6093):163-6.

Saiki, R. K., Walsh, P. S., Levenson, C. H. and Erlich, H. A. (1989). Genetic analysis of amplified DNA with immobilized sequence-specific oligonucleotide probes. Proc Natl Acad Sci U S A 86(16):6230-4.

Sanger, F., Nicklen, S. and Coulson, A. R. (1977). DNA sequencing with chain-terminating inhibitors. Proc Natl Acad Sci U S A 74(12):5463-7.

Sauer, S., Lechner, D., Berlin, K., Lehrach, H., Escary, J. L., Fox, N. and Gut, I. G. (2000). A novel procedure for efficient genotyping of single nucleotide polymor-phisms. Nucleic Acids Res 28(5):e13.

Sauer, S., Gelfand, D. H., Boussicault, F., Bauer, K., Reichert, F. and Gut, I. G. (2002). Facile method for automated genotyping of single nucleotide polymor-phisms by mass spectrometry. Nucleic Acids Res 30(5):e22.

Schena, M., Shalon, D., Davis, R. W. and Brown, P. O. (1995). Quantitative monitor-ing of gene expression patterns with a complementary DNA microarray. Science270(5235):467-70.

Schillaci, G., Verdecchia, P., Porcellati, C., Cuccurullo, O., Cosco, C. and Perticone, F. (2000). Continuous relation between left ventricular mass and cardiovascular risk in essential hypertension. Hypertension 35(2):580-6.

Seidman, E. G. (2003). Clinical use and practical application of TPMT enzyme and 6-mercaptopurine metabolite monitoring in IBD. Rev Gastroenterol Disord 3 Suppl 1(S30-8.

68

Serafinowski, P. J. and Garland, P. B. (2003). Novel photoacid generators for photodi-rected oligonucleotide synthesis. J Am Chem Soc 125(4):962-5.

Shifman, S., Kuypers, J., Kokoris, M., Yakir, B. and Darvasi, A. (2003). Linkage dis-equilibrium patterns of the human genome across populations. Hum Mol Genet12(7):771-6.

Shuber, A. P., Grondin, V. J. and Klinger, K. W. (1995). A simplified procedure for developing multiplex PCRs. Genome Res 5(5):488-93.

Singh-Gasson, S., Green, R. D., Yue, Y., Nelson, C., Blattner, F., Sussman, M. R. and Cerrina, F. (1999). Maskless fabrication of light-directed oligonucleotide microar-rays using a digital micromirror array. Nat Biotechnol 17(10):974-8.

Smith, R. A., Ho, P. J., Clegg, J. B., Kidd, J. R. and Thein, S. L. (1998). Recombina-tion breakpoints in the human beta-globin gene cluster. Blood 92(11):4415-21.

Sofowora, G. G., Dishy, V., Muszkat, M., Xie, H. G., Kim, R. B., Harris, P. A., Prasad, H. C., Byrne, D. W., Nair, U. B., Wood, A. J., et al. (2003). A common beta1-adrenergic receptor polymorphism (Arg389Gly) affects blood pressure re-sponse to beta-blockade. Clin Pharmacol Ther 73(4):366-71.

Sorek, R., Shamir, R. and Ast, G. (2004). How prevalent is functional alternative splicing in the human genome? Trends Genet 20(2):68-71.

Sosnowski, R. G., Tu, E., Butler, W. F., O'Connell, J. P. and Heller, M. J. (1997). Rapid determination of single base mismatch mutations in DNA hybrids by direct electric field control. Proc Natl Acad Sci U S A 94(4):1119-23.

Southern, E. M., Maskos, U. and Elder, J. K. (1992). Analyzing and comparing nucleic acid sequences by hybridization to arrays of oligonucleotides: evaluation using ex-perimental models. Genomics 13(4):1008-17.

Southern, E. M. (1996). DNA chips: analysing sequence by hybridization to oligonu-cleotides on a large scale. Trends Genet 12(3):110-5.

Staessen, J. A., Wang, J., Bianchi, G. and Birkenhager, W. H. (2003). Essential hyper-tension. Lancet 361(9369):1629-41.

Stenman, J., Finne, P., Stahls, A., Grenman, R., Stenman, U. H., Palotie, A. and Or-pana, A. (1999). Accurate determination of relative messenger RNA levels by RT-PCR. Nat Biotechnol 17(7):720-2.

Stumpf, M. P. and Goldstein, D. B. (2003). Demography, recombination hotspot intensity, and the block structure of linkage disequilibrium. Curr Biol 13(1):1-8.

Suomalainen, A., Kollmann, P., Octave, J. N., Soderlund, H. and Syvanen, A. C. (1993). Quantification of mitochondrial DNA carrying the tRNA(8344Lys) point mutation in myoclonus epilepsy and ragged-red-fiber disease. Eur J Hum Genet1(1):88-95.

Syvanen, A. C., Aalto-Setala, K., Harju, L., Kontula, K. and Soderlund, H. (1990). A primer-guided nucleotide incorporation assay in the genotyping of apolipoprotein E. Genomics 8(4):684-92.

Syvanen, A. C., Sajantila, A. and Lukka, M. (1993). Identification of individuals by analysis of biallelic DNA markers, using PCR and solid-phase minisequencing. Am J Hum Genet 52(1):46-59.

Tishkoff, S. A. and Verrelli, B. C. (2003). Role of evolutionary history on haplotype block structure in the human genome: implications for disease mapping. Curr Opin Genet Dev 13(6):569-75.

Tsuchihashi, Z. and Dracopoli, N. C. (2002). Progress in high throughput SNP geno-typing methods. Pharmacogenomics J 2(2):103-10.

Turner, S. T., Schwartz, G. L., Chapman, A. B., Hall, W. D. and Boerwinkle, E. (2001). Antihypertensive pharmacogenetics: getting the right drug into the right patient. J Hypertens 19(1):1-11.

69

Turner, S. T. and Boerwinkle, E. (2003). Genetics of blood pressure, hypertensive complications, and antihypertensive drug responses. Pharmacogenomics 4(1):53-65.

Wadelius, M., Sorlin, K., Wallerman, O., Karlsson, J., Yue, Q. Y., Magnusson, P. K., Wadelius, C. and Melhus, H. (2004). Warfarin sensitivity related to CYP2C9, CYP3A5, ABCB1 (MDR1) and other factors. Pharmacogenomics J 4(1):40-8.

Wall, J. D. and Pritchard, J. K. (2003). Assessing the performance of the haplotype block model of linkage disequilibrium. Am J Hum Genet 73(3):502-15.

Wallace, R. B., Shaffer, J., Murphy, R. F., Bonner, J., Hirose, T. and Itakura, K. (1979). Hybridization of synthetic oligodeoxyribonucleotides to phi chi 174 DNA: the effect of single base pair mismatch. Nucleic Acids Res 6(11):3543-57.

Walsh, M. K., Wang, X. and Weimer, B. C. (2001). Optimizing the immobilization of single-stranded DNA onto glass beads. J Biochem Biophys Methods 47(3):221-31.

Van Eerdewegh, P., Little, R. D., Dupuis, J., Del Mastro, R. G., Falls, K., Simon, J., Torrey, D., Pandit, S., McKenny, J., Braunschweiger, K., et al. (2002). Association of the ADAM33 gene with asthma and bronchial hyperresponsiveness. Nature418(6896):426-30.

van Eijk, M. J., Broekhof, J. L., van der Poel, H. J., Hogers, R. C., Schneiders, H., Kamerbeek, J., Verstege, E., van Aart, J. W., Geerlings, H., Buntjer, J. B., et al.(2004). SNPWave: a flexible multiplexed SNP genotyping technology. Nucleic Ac-ids Res 32(4):e47.

Van Ness, J., Kalbfleisch, S., Petrie, C. R., Reed, M. W., Tabone, J. C. and Vermeulen, N. M. (1991). A versatile solid support system for oligodeoxynucleotide probe-based hybridization assays. Nucleic Acids Res 19(12):3345-50.

Watson, J. D. and Crick, F. H. (1953). Genetical implications of the structure of de-oxyribonucleic acid. Nature 171(4361):964-7.

Weiler, J., Gausepohl, H., Hauser, N., Jensen, O. N. and Hoheisel, J. D. (1997). Hy-bridisation based DNA screening on peptide nucleic acid (PNA) oligomer arrays. Nucleic Acids Res 25(14):2792-9.

Weinshilboum, R. (2003). Inheritance and drug response. N Engl J Med 348(6):529-37.

Venter, J. C., Adams, M. D., Myers, E. W., Li, P. W., Mural, R. J., Sutton, G. G., Smith, H. O., Yandell, M., Evans, C. A., Holt, R. A., et al. (2001). The sequence of the human genome. Science 291(5507):1304-51.

WHO. (2003). 2003 World Health Organization (WHO)/International Society of Hypertension (ISH) statement on management of hypertension. J Hypertens21(11):1983-1992.

Vieux, E. F., Kwok, P. Y. and Miller, R. D. (2002). Primer design for PCR and se-quencing in high-throughput analysis of SNPs. Biotechniques Suppl(28-30, 32.

Yan, H., Yuan, W., Velculescu, V. E., Vogelstein, B. and Kinzler, K. W. (2002). Alle-lic variation in human gene expression. Science 297(5584):1143.

Yuryev, A., Huang, J., Pohl, M., Patch, R., Watson, F., Bell, P., Donaldson, M., Phil-lips, M. S. and Boyce-Jacino, M. T. (2002). Predicting the success of primer exten-sion genotyping assays using statistical modeling. Nucleic Acids Res 30(23):e131.

Acta Universitatis UpsaliensisComprehensive Summaries of Uppsala Dissertations

from the Faculty of MedicineEditor: The Dean of the Faculty of Medicine

Distribution:Uppsala University Library

Box 510, SE-751 20 Uppsala, Swedenwww.uu.se, acta@ub.uu.se

ISSN 0282-7476ISBN 91-554-5937-4

A doctoral dissertation from the Faculty of Medicine, Uppsala University,is usually a summary of a number of papers. A few copies of the completedissertation are kept at major Swedish research libraries, while the sum-mary alone is distributed internationally through the series Comprehen-sive Summaries of Uppsala Dissertations from the Faculty of Medicine.(Prior to October, 1985, the series was published under the title “Abstracts ofUppsala Dissertations from the Faculty of Medicine”.)