Transition to college and CAO Fiona Casey Schools’ Liaison Officer NUI Maynooth.
Justin Blumenstiel (1) Fiona He (2) Casey M....
Transcript of Justin Blumenstiel (1) Fiona He (2) Casey M....
Justin Blumenstiel (1)Fiona He (2)
Casey M. Bergman (2)
(1) Dept. of Ecology & Evolutionary Biology, University of Kansas(2) Faculty of Life Sciences, University of Manchester
An age-of-allele test of neutrality for transposable element insertions
A brief introduction to models of TE evolution
• Selfish DNA sequences, intra-genomic parasites
• Transposition rates >> excision rates
• Equilibrium maintained by transposition-selection balance
• Mechanism of negative selection is debated
• TE insertions at low frequency because of negative selection
• Recent genomic evidence casts doubt on the assumption of equilibrium and inference of negative selection
Petrov & Hartl (1998) Mol. Biol. Evol. 15:293-302
Estimating the age of ‘pseudogene-like’ retrotransposon insertion alleles
D. mel - D. sim speciation
Bergman & Bensasson (2007) PNAS 104:11340-5
a_in
vader2
_6
b_m
icro
pia
_4
c_T
abor_
3
d_17.6
_11
e_S
talk
er_
4
f_ro
ver_
3
g_flea_16
h_copia
_28
i_m
dg3_10
j_ro
o_86
k_T
ranspac_4
l_opus_16
m_blo
od_22
n_412_24
o_B
urd
ock_13
p_div
er_
9
q_T
irant_
20
r_jo
ckey2_7
s_H
ele
na_7
t_C
r1a_36
u_baggin
s_6
v_G
4_10
w_D
oc3_7
x_G
5_8
y_B
S_15
z_Juan_9
zz_D
oc_53
0.00
0.02
0.04
0.06
0.08
0.10
0.12D
iverg
ence (
sub/s
ite)
0
1.80
3.60
5.41
7.21
9.01
10.81
Age (
Mya)
Retrotransposon demographics in D. melanogaster
Bartolome et al. (2009) Genome Biology 10:R22
Horizontal transfer of D. melanogaster TE families
silent site divergence
Project aims
• Develop a non-equilibrium model of neutral TE evolution that relaxes the assumption of a constant TE insertion rate.
• Obtain allele frequency data for a large sample of TEs in ancestral and derived populations of D. melanogaster.
• Test whether observed TE allele frequencies are consistent with ages of TE insertion estimated from genomic data to infer forces controlling TE evolution.
An age-of-allele model for TE insertions
• Question: what is the probability that an allele of age t is present in i copies in a sample of n chromosomes?
• Calculate probability of i descendants from a single ancestor given j ancestors (Feller 1957)
• Calculate probability of j ancestors at time t under standard neutral model (Tavare 1984)
• Calculate probability of insertion at time t given s substitutions in a fragment of length l under Poisson process (Bayes 1763)
Allele frequency data for TE insertions
• 190 loci (90 LTR and 100 non-LTR)
• 2 PCR per loci per strain (TE+flank / L+R flanking regions)
• 12 strains from 2 populations - Zimbabwe (from Stephan Lab) & North Carolina (from Mackay Lab)
• Insertion in genomic sequence is included as 13th allele to account for ascertainment bias
• Individual strain allele frequency data consistent with pooled strain allele frequency data from Gonzalez et al. (2008)
lower frequencythan expected
higher frequencythan expected
Fit of expected allele frequency under neutral model to observed frequency in North Carolina
0 50 100 150
-50
510
rank difference
observed-expected
Expected allele frequency fits observed allele frequency over a wide range of ages
-8 -7 -6 -5 -4 -3
-50
510
log(subs/site)
observed-expected
Preliminary observations
• Majority of TE insertions in North Carolina are at or close to expected frequency given age since insertion under neutrality
• Some loci deviate strongly from predicted frequency and may reflect loci under positive and negative selection
• Model accurately predicts observed allele frequency over wide range of insertion ages
• Model parameterized with current estimate of African population size leads to poor predictions but yields better fit with ancestral population size
Ongoing and Future Work
• Use model to generate maximum likelihood estimate of Ne under assumption that insertion alleles are neutral
• Analysis of the fit of model to data according to:
- TE class
- TE family
- X vs. autosome
- recombination
• Resolution of best summary statistic(s) to assess global fit of the model to the data
Justin Blumenstiel Fiona He
Nicolas SvetecWolfgang Stephan
Acknowledgements