Reconstructing a herbivore’s diet using a novel rbcL DNA ... · they consume. This is in contrast...

Invited Review

SPECIAL ISSUE Interactions between White-Tailed Deer andInvasive Plants in North American Forests

Reconstructing a herbivorersquos diet using a novel rbcL DNAmini-barcode for plants

David L Erickson1 Elizabeth Reed1 Padmini Ramachandran1 Norman A Bourg2William J McShea3 and Andrea Ottesen1

1 Molecular Methods and Subtyping Branch Division of Microbiology Centre for Food Safety and Applied Nutrition (CFSAN) FDA5001 Campus Drive College Park ML 20740 USA2 US Geological Survey National Research Program - Eastern Branch 12201 Sunrise Valley Drive MS 430 Reston VA 20192 USA3 Smithsonian Conservation Biology Institute National Zoological Park 1500 Remount Road Front Royal VA 22630 USA

Received 26 August 2016 Editorial decision 28 March 2017 Accepted 13 April 2017 Published 21 April 2017

Guest Editor David Gorchov

Citation Erickson DL Reed E Ramachandran P Bourg NA McShea WJ Ottesen A 2017 Reconstructing a herbivorersquos diet using anovel rbcL DNA mini-barcode for plants AoB PLANTS 9 plx015 doi101093aobplaplx015

Abstract Next Generation Sequencing and the application of metagenomic analyses can be used to answer ques-tions about animal diet choice and study the consequences of selective foraging by herbivores The quantification ofherbivore diet choice with respect to native versus exotic plant species is particularly relevant given concerns of invasivespecies establishment and their effects on ecosystems While increased abundance of white-tailed deer (Odocoileus vir-ginianus) appears to correlate with increased incidence of invasive plant species data supporting a causal link is scarceWe used a metabarcoding approach (PCR amplicons of the plant rbcL gene) to survey the diet of white-tailed deer (fecalsamples) from a forested site in Warren County Virginia with a comprehensive plant species inventory and correspond-ing reference collection of plant barcode and chloroplast sequences We sampled fecal pellet piles and extracted DNAfrom 12 individual deer in October 2014 These samples were compared to a reference DNA library of plant species col-lected within the study area For 72 of the amplicons we were able to assign taxonomy at the species level whichprovides for the first timemdashsufficient taxonomic resolution to quantify the relative frequency at which native and exoticplant species are being consumed by white-tailed deer For each of the 12 individual deer we collected three subsamplesfrom the same fecal sample resulting in sequencing 36 total samples Using Qiime we quantified the plant DNA foundin all 36 samples and found that variance within samples was less than variance between samples (Ffrac14173 Pfrac140004)indicating additional subsamples may not be necessary Species level diversity ranged from 60 to 93 OTUs per individualand nearly 70 of all plant sequences recovered were from native plant species The number of species detected did re-duce significantly (range 4ndash12) when we excluded species whose OTU composedlt1 of each samplersquos total Whencompared to the abundance of native and non-natives plants inventoried in the local community our results supportthe observation that white-tailed deer have strong foraging preferences but these preferences were not consistent forspecies in either class Deer forage behaviour may favour some exotic species but not all

Keywords Deer diet exotic plants fecal DNA forest metagenomics North America oaks white-tailed deer

Corresponding authorrsquos e-mail address mcsheawsiedu

Published by Oxford University Press on behalf of the Annals of Botany Company 2017 This work is written by US Governmentemployees and is in the public domain in the US

AoB PLANTS wwwaobplantsoxfordjournalsorg VC The Authors 2017 100

Introduction

Trophic interactions among species are thought to be acritical regulatory mechanism affecting species fitnessand community composition (Paine 1966 Shears andBabcock 2003) These interactions can take many formsthe most familiar of which include predator-prey (Letnicet al 2009) hostndashpathogen (Hajishengallis et al 2012)and plantndashherbivore (Creed 2000) These interactions ap-pear to be part of a broader cascade of interaction ef-fects all with significant impacts on individual speciesand the communities they inhabit (Cumplido et al 2016)While invertebrate herbivory has been widely studiedand implicated in structuring plant communities (Janzen1970 Maron and Crone 2006 McShea et al 2007) largeherbivores can also shift plant community compositionin both forests (Frerker et al 2014) and grasslands(Augustine and McNaughton 1998) although their role isless well understood Improved quantification of ungu-late diet will greatly contribute to our understanding ofhow foraging behaviour sustains or degrades naturalcommunities and importantly how the removal of spe-cies from communities through extinction and extirpa-tion may have a cascading effect on naturalcommunities

Increasingly molecular genetic tools to quantify tro-phic interactions have been explored from a technicalperspective (Hibert et al 2013 Ficetola et al 2016)These studies follow the pathway of metagenomic anal-yses first pioneered in environmental and human micro-biome studies of bacteria (Lane et al 1985 Hugenholtzet al 1998) using PCR amplicon data from DNA samplescontaining a mixture of species and large referencedatabases for those genes notably 16S in microbialcommunities (Baker and Banfield 2003 Sogin et al 2006Turnbaugh et al 2009) In animal studies the CO1 genethe official DNA barcode in animals has been utilized(Deagle et al 2009 Leray et al 2013) while in herbivoresdifferent markers have been used (Pompanon et al2012 Garcıa-Robledo et al 2013) Several program pack-ages have been developed to help streamline the pro-cess of taxonomic identification of the sequences inmixed samples (such as fecal samples containing plantDNA) notably Qiime (Caporaso et al 2010) Most studiesin eukaryotes have focused on accurately measuring di-versity (Leray and Knowlton 2015 Goldberg et al 2015)and have begun to quantify interactions between species(Murray et al 2011) There have been comparativelyfewer studies that directly engage in hypothesis testing(Leray et al 2015 Kartzinel et al 2015 Soininen et al2009) compared to bacterial metagenomics (Blaser2014 Thomsen and Willerslev 2015) But numerous re-views have highlighted the utility of these methods in

ecology (Valentini et al 2009 Bourlat et al 2013 Clare2014) and increasing numbers of reports implementingmetagenomic methods are being published

Studies that use metagenomics to diagnose interac-tions between different trophic levels in natural commu-nities are rare (eg Kartzinel et al 2015 Murray et al2011 Soininen et al 2009) particularly compared to thewidespread use of metagenomics to quantify organismaldiversity in an array of environments (Taberlet et al2012) Nonetheless using metagenomics in trophic in-teractions has been well reviewed (Valentini et al 2009Bourlat et al 2013 Clare 2014) Most metagenomic stud-ies of inter-trophic interactions have focused on predatordiets (eg Deagle et al 2009 2010 Bourlat et al 2013Leray et al 2013 2015) with few focusing on herbivorediet (Soininen et al 2009 Kartzinel et al 2015) Yet thepower that metagenomics offers in revealing thestrength of trophic interactions suggests that this will be-come an increasingly common tool particularly forquantifying animal diet when the animalsrsquo feedingbehaviour is not directly observable Deducing herbivorediet is relatively simple as the DNA of the herbivore itselfwill not confound the bioinformatic analysis of the plantsthey consume This is in contrast to carnivore diets wherethe DNA of the carnivore may be PCR amplified alongwith the diet of its prey (Leray et al 2013) Yet metage-nomic studies investigating trophic interactions amongherbivores and their plant community have also beenrare especially relative to the role that herbivory plays inshaping natural plant communities (Cote et al 20042014 McShea 2012 Holm et al 2013 Frerker et al2014)

Historically observations of herbivore diet have largelytaken three forms direct observation (Holechek et al1982) exclusion experiments (Bowers 1993 Parker et al2006) and micro-histological analysis of animal scats(Alipayo et al 1992 Kowalczyk et al 2011) All of thesemethods have various limitations but the use of geneticsto deduce animal diet from feces was challenging be-cause the traditional sequencing technology could onlydeduce one species at a time with accuracy thereby be-ing limited to use only with highly specialized herbivores(Garcıa-Robledo et al 2013) However for generalist her-bivores with broad diets the recent advances in mas-sively parallel Next-Generation Sequencing methodsenables us to simultaneously sequence DNA from allplant species in a mixture (Soininen et al 2009)

In the forests of North America exotic insects andplants have caused significant ecosystem alteration anda body of theory has developed around the mechanismsby which some exotic species become invasive (Sanderset al 2003) There are multiple hypotheses for invasivespecies success that rely on preferential consumption of

Erickson et al mdash Deer diet and fecal DNA

002 AoB PLANTS wwwaobplantsoxfordjournalsorg VC The Authors 2017

native plant species For example the enemy-release hy-pothesis (Keane and Crawley 2002) proposes that nativespecies have the full complement of natural enemies(herbivores) but invasive species do not likewise thenovel defense hypothesis postulates exotic plants offernovel and therefore more effective plant defensesagainst herbivores (Enge et al 2012 Macel et al 2014)Deer are known to feed selectively on forest plants(Rooney 2001) but direct observation of deer behaviouris difficult and recent experiments have instead used acommon garden approach to determine preference(Averill et al 2016) But such experiments have limitedscope due to the small number of plants that can becompared relative to the natural diet breadth of mostherbivores Effective means to deduce the diet of deer inwild populations would be critical to understanding theirimpact on invasive plants

White-tailed deer (Odocoileus virginianus) are a wide-spread and often important herbivore in eastern decidu-ous forests of North America (McShea 2012) Throughoutmost of eastern North America present densities of deerexceed historical levels (McCabe and McCabe 1997)These high densities have been implicated in changingthe composition of herbaceous and woody plant com-munities in many forest ecosystems (Vavra et al 2007)shifting the chemical composition of soils and succes-sional pathways (Pastor and Naiman 1992) and facilitat-ing the invasion of exotic plant species (Shen et al 2016)However although the abundance of deer has corre-sponded with increased rates of invasive plant abun-dance the mechanisms behind this relationship remainuncertain (Averill et al 2016) Using metagenomics tohelp determine the forage species and their relative fre-quency in wild populations is of major economic andconservation consequence

The availability of reference sequence data for plantsmay be one limitation that prevents the more wide-spread application of metagenomics to the study of her-bivore diets The Smithsonian Conservation BiologyInstitute (SCBI) is the site of one of the few natural habi-tats where most plants within a study area have been se-quenced to obtain a genetic fingerprint through DNAbarcoding (Valentini 2009 Erickson et al 2014) In thisstudy we use a metagenomic approach of applying aPCR-based mini-barcode marker in conjunction with theanalysis package Qiime (Caporaso et al 2010) to addressseveral questions affecting the diagnosis of animal dietsin natural environments namely (i) what is the efficacyof using DNA recovered from fecal samples to producePCR products that can distinguish plant species (ii) Dorepeat samples of the same fecal pile indicate caution isneeded when sampling fecal DNA for diet information(iii) Can a single locus rbcL barcode distinguish species in

a reference DNA library and be used to accurately assignfecal DNA to tree shrub and herb species and (iv) canobservational and relative abundance data from thenumber of operational taxonomic units (OTUs) and bar-coding be used to infer herbivore diet selection for oragainst invasive plants

Methods

Study organisms and study site

This study focused on the diet of white-tailed deer(Odocoileus virginianus) at the SCBI in Front Royal VAadjacent to the northern end of Shenandoah NationalPark Within SCBI our sampling focused around a 26 haLarge Forest Dynamics Plot (LFDP) because the vascularplant flora was exhaustively sampled and genotyped atthe accepted DNA barcode loci (CBOL Plant WorkingGroup 2009) Those DNA barcode reference plant se-quences were deposited in GenBank and are availableunder accessions KP642769ndashKP644136 Vouchers foreach plant included in the reference set are stored at theNational Museum of Natural Historyrsquos herbarium We listplant species as native introduced and invasive The ter-minology of exotic plants in problematic (Richardsonet al 2000 Valery et al 2008 Pereyra 2016) We differ-entiate between two types of exotic plants introducedspecies that persist within native plant communities andinvasive species that are undergoing rapid spread inrange and habitat associations and tend to rapidly in-crease in numbers once established We referenced theUSDA to identify introduced species (USDA NRCS 2017)and the state of Virginia to identify those introduced spe-cies that are considered invasive in our region (VA DeptCons and Rec 2015) When plants were only identifiedto genus and the genus contained a mixture introducedand native species we designated the genus accordingto the most abundant species For example the genusAgrotis Festuca and Poa were considered introducedand genus Acer was considered native

Sampling of the vascular plant flora was accomplishedby collecting leaf material and voucher specimens for allwoody vascular plant speciesgt1 cm DBH vines andherbs present on the LFDP to produce a complete DNAbarcode library for this representative mature secondaryforest This was achieved with a three-pronged surveyapproach encompassing a subset of 330 one m2 woodyseedling plots in place on the LFDP for systematic sam-pling intensive collection within a 4-ha long-term deerexclosure located within the LFDP and walking searchesfor unrecorded species conducted across the LFDPWhenever possible at least three individuals of a givenspecies were sampled for leaf tissue collected and

Erickson et al ndash Deer diet and fecal DNA


subsequently genotyped at the four sequence regions ofrbcL matK trnH-psbA and ITS (Kress et al 2009) TheDNA barcodes from these collections were combinedwith the previously generated barcodes from 66 speciesof woody plants vines and herbs collected previously onthe LFDP as well as the 62 species of woody trees andshrubs sampled during the initial plot census (Bourget al 2013 Erickson et al 2014)

The deer diet as estimated from the fecal sampleswas compared to the relative abundance of plants withinthe study area In JuneJuly 2009 we sampled 160 plotswithin the study area to estimate species richness andrelative abundance for the site We selected 40 randompoints within the mature forest and arrayed four 1 1 mplots around each selected point At each plot we identi-fied all woody and herbaceous stems (up to 2 m height)to species and recorded abundances for all species (num-ber of individuals for woody and forb species and percentcover for graminoids and ferns) For this paper we usedthe presence of species in each 1 1 m plot to calculatethe proportion of plots containing the species andthereby the relative abundance of each species detected

Fecal sampling and DNA isolation

Twelve deer fecal samples were collected in the field onOctober 10ndash15 2014 and collections were within for-ested areas within 200 m of the LFDP Samples were onlycollected from obviously lsquofreshrsquo fecal piles as evidencedby no degradation of pellet form and all pellets moist onthe surface The pellets were handled with sterile glovesand 10 g were placed in a marked ziplock bag and ho-mogenized within the bag The steps by which thesesamples were processed is outlined in Fig 1 Upon re-turning to lab three samples 01 g each of fecal mate-rial were then subsampled from each bag and placedinto 2 mL polystyrene tubes with screw caps and O-ringseals to prevent leakage and contamination resulting in36 total samples Each collected sample was mixed with800 mL of Zymo XpeditionTM LysisStabilization Solutionthen thoroughly mixed by vortexing with the Lysis bufferon a laboratory mixer to ensure complete suspension ofthe buffer and sample Following mixing samples werefrozen and stored at 20 C in Front Royal Virginia untilDNA isolation in the laboratory

DNA isolation was conducted in the Laboratories forAnalytical Biology at the Smithsonian Institution (LAB-SI) Frozen samples were transferred to LAB-SI and keptat 20 C For DNA extraction the samples were thawedand then processed utilizing the Zymo ZR Fecal DNAMiniPrep kit to recover and purify DNA After thawing the36 samples were immediately centrifuged at maximumspeed (14 000 rpm) in a mini-centrifuge for 5 min after

which 400 mL of supernatant was used for downstreamDNA isolation Use of the kit involved separating solidsfrom the lysis buffer binding of the DNA to silica in a spincolumn and elution in 100 mL of 10 mM Tris-1 mM EDTA(pH 80) buffer The eluted DNA was quantified to ensureits capture and then stored at 20 C until used in PCRWe employed a negative control composed of sterilewater that was processed in parallel to the 12 fecal sam-ples The supernatant for the negative control was thenused in both the PCR and WGS sequencing (see below)

To ensure that each of fecal samples were derivedfrom different individuals we used two microsatellitesdeveloped for genetic fingerprinting of white-tailed deer(Anderson et al 2002) using markers denoted as N andQ PCR amplification using the DNA recovered from fecalsamples was conducted for all 36 DNA samples followingAnderson et al (2002) and the different size alleles foreach animal were resolved using an Agilent TapeStationwith high sensitivity tapes

PCR

A mini-barcode approach was chosen for PCR amplifica-tion of plant DNA recovered from the 36 fecal samples

Figure 1 Flow chart of procedures for collecting extracting andquantifying DNA used to identify deer forage species and deer iden-tification for 12 samples (1) Samples were collected within a 1 km2

at Smithsonianrsquos Conservation Biology in Front Royal VA in October2014 (2) The DNA extraction was conducted using Zymo Fecal DNAMini-prep Kit (3a) Known microsatellite markers were used verifyeach of the 12 samples were from separate individuals (3b) In par-allel to the PCR based mini-barcode approach WGS data was gener-ated and analysed using five whole chloroplast sequences in CLCBio (3c) PCR was conducted using primers designed for mini-Barcode rbcl region (4a) Sequencing of amplicons was performedon an Illumina MiSeq V2 (4b) Voucher DNA Barcode samples fromplant species collected in region were obtained from GenBank (5)Data quality control was performed in CLC Bio Genomics Workbench(ver 74) while bioinformatics analysis where sequences were as-signed taxonomic identity were performed using Qiime (6) Primarystatistics contrasting animal diet included PERMANOVA and ANOSIMwhich utilized OTU tables from Qiime analysis



The mini-barcodes were derived from the standard rbcLplant DNA barcode region using the accepted reverse pri-mer from standard DNA barcodes and a forward primerbuilt following multiple sequence alignment of over 500rbcL sequences derived from 30 separate orders (includ-ing the Polypodiales Pinales and the remainder from themost speciose clades of Magnoliophyta) and identifica-tion of a sub-region with no mismatches in the criticalfirst five bases of the 30 end for the set of 500 sequencesused in our design While this analysis did not explicitlyquantify potential primer bias which is possible se-quence alignment and in-silico testing using completechloroplast genomes for species from AsteraceaeFagales Malpighiales and Poaceae suggested limited pri-mer mismatch The new forward locus specific primer se-quence we used was 50-CTTACCAGYCTTGATCGTTACAAAGG-30 and the reverse primer sequence which is the re-verse primer for the official rbcL DNA barcode region is50-GTAAAATCAAGTCCACCRCG-30 (Kress and Erickson2007) which is the accepted and widely implementedstandard primer used in rbcL plant barcoding (CBOLPlant Working Group 2009) The amplicon size from thisprimer pair is a constant 379 base pairs across our align-ment of the 500 rbcL sequences Primers were synthe-sized as fusion primers with the Illumina Nexteraadaptors at the 50 end following the 16S MetagenomicProtocol as Forward fusion primer 50-TCGTCGGCAGCGTCAGATGTGTATAAGAGACAG-[locus specific se-quence] and Reverse fusion primer 50-GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAG-[locus specificsequence]

Extracted DNA was used for conducting PCR to amplifyrbcL minibarcode All DNA samples were subjected toidentical PCR conditions The reaction cocktail consistedof 125 mL of 2 EmeraldAmp MAX PCR Master Mix 01mL each of 100 mM forward and reverse primer 1 mL ofDNA and 113 mL of ultrapure water Thermocycling wasconducted in a ABI 2720 with the program 1 cycle of 95C 4 min 35 cycles of 94 C for 20 s 55 C for 30 s 72 Cfor 1 min 1 final extension of 72 C for 5 min and a holdat 9 C until the reaction was removed Each DNA samplehad its PCR reaction repeated once and the two reac-tions were pooled in equal concentration as measuredusing a Qubit fluorometer prior to PCR cleanup The neg-ative control was used in this protocol to test for howlaboratory processing affected results

The Next Generation Sequencing library for the 36 fe-cal pellet PCR samples followed the Illumina 16S proto-col (see Illumina 2013) Amplicons from the rbcL PCRwere verified through gel electrophoresis on an AgilentTapeStation then purified with Ampure beads (08 vol-ume of beads to PCR volume) and washed with 80 Ethanol twice DNA was eluted from beads by mixing

525 mL of 10 mM Tris (pH 85) buffer with the beads andrecovering 50 mL of clean DNA solution following separa-tion of the beads with a magnet Clean up was per-formed in a 96 well plate using a magnet plate fromAmbion The cleaned PCR product was then used in asecond PCR which added a unique combination ofNextera XT indexes (Illumina catalog number FC-131-1002) for each of the 36 samples For each index PCRamplification we used 5 mL of cleaned PCR product 5 mLeach of the Nextera XT index primers (S5XX and N7XX re-spectively) 25 mL of 2 EmeraldAmp MAX PCR MasterMix and 10 mL of H2O Each reaction conducted on anABI 2720 using the program one cycle of 95 C 3 min 8cycles of 95 C 30 s 55 C 30 s 72 C 30 s 1 final exten-sion of 72 C 5 min followed by a hold at 9 C until thereaction was removed The index PCR was cleaned usinga 112 volume of Ampure beads (56 mL of Ampurebeads mixed with 50 mL PCR) and washed twice with80 Ethanol while on a magnet stand Clean index PCRwas eluted from beads using 275 mL of 10 mM Tris (pH85) buffer with 25 mL of solution recovered

Each cleaned index PCR was quantified using fluoro-metric assay with a Qubit 30 system using 3 mL of DNAin each measurement Products were pooled in equalmolar concentrations to a final 40 mM solution contain-ing all 36 samples

Evaluation of rbcLmini-barcode locus

To evaluate the ability of the rbcL-mini-barcode locusused to correctly identify plant species from specimenscollected at SCBI we tested different fractions of thestandard rbcL barcode locus relative to the mini-barcode This involved alignment of all rbcL referencesequences for SCBI (nfrac14234) including at least two indi-viduals per species That alignment was then trimmed todifferent lengths 100 200 300 and 400 bp We used asliding window that was used to extract 100 bp seg-ments from along the length of the reference DNA bar-code that is 540 bp in length The sliding window was in20 bp increments such that there were 23 different100 bp segments 18 different 300 bp segments eightdifferent 400 bp segments and three different 500 bpsegments Each of these segments which represented asampling of the aligned reference database was used inthe SpeciesIdentifier module in TaxonDNA v18 (Meieret al 2006) to quantify rates of identification to differenttaxonomic levels For each sequence segment tested werecorded the rate of correct identification to species ge-nus and family level The average value for identificationto each of the three taxonomic levels was then plotted(Fig 2) The values for species level identification for themini-barcode and for the full length DNA barcode were



determined using TaxonDNA and plotted with the resultsfor the sliding window of difference size selections

Sequencing and post-sequence processing

The 40 mM sample of pooled indexed PCR was diluted to4 pM and sequenced using an Illumina MiSeq with a500 bp V2 kit with 251 cycles Sequences produced onthe MiSeq had adaptor and index sequences trimmed onthe instrument by bcl2fastq Conversion Software (v184httpsupportilluminacom) with each of the 36 fecalDNA samples resulting in separate forward and reversesequences For each of the 36 fecal DNA samples thesoftware package CLC Genomics Workbench (ver 74)was used to trim sequences remove locus specific PCRprimer sequences and merge the forward and reversestrands All sequences were quality filtered and trimmedbefore merging into contigs The software was used to re-move all sequences containing ambiguous bases andtrimmed all sequences at both ends using quality scoresto a 003 limit (equivalent to a Q value of 30) with furtherdetail on the trimming algorithm found here (httpresourcesqiagenbioinformaticscommanualsclcgenomicsworkbench801indexphpmanual frac14 Quality_trimminghtml) Only sequences that could be merged were used insubsequent analyses and we used the following parame-ters within CLC Bio Genomics Workbench (ver 74) mis-match cost frac14 2 Gap cost frac14 4 Max unaligned frac14 0Minimum score frac14 0 again for more description of theimplementation of the parameters see httpresourcesqiagenbioinformaticscommanualsclcgenomicsworkbench801indexphpmanual frac14 Merge_overlapping_pairshtml For each of the 36 samples we used CLC BioGenomics Workbench (ver 74) to randomly select100 000 merged sequences for use in downstream analy-sis The 100 000 randomly selected sequences were thenexported in FASTA format for downstream use in assign-ment of plant identity

Metagenomic analysis of fecal plant DNA

We define an OTU as a DNA sequence cluster that corre-sponds to a taxonomic group either Family Genus orSpecies (Blaxter et al 2005) Whether an OTU can be as-signed to a specific taxonomic level depends on the reso-lution of the mini-barcode sequence We assignedspecies identifications to the OTU extracted from the fe-cal DNA using Qiime (Caporaso et al 2010) in combina-tion with the reference sequences for the LFDP toidentify the species of plants consumed by the white-tailed deer We followed the 454 SOP (see httpqiimeorgtutorialstutorialhtml) which involved removal ofchimera and OTU picking assignment of taxonomy foreach OTU based on the reference database and output

of a BIOM table that contained all taxonomic and abun-dance data The BIOM table was converted usingassign_taxonomypy to list the taxonomy of identifiedOTU and the number of sequences that were in eachOTU We implemented the closed reference OTU algo-rithm focusing only on assignment to the plant speciesfound in our reference library for SCBI through theldquopick_closed_reference_otuspyrdquo option in Qiime Weused uclust as the cluster algorithm for generating OTUat a 97 similarity threshold The reference sequenceswere only for the rbcL gene and were trimmed to thesame constant 379 bp corresponding to our mini-barcode and pared such that only a single sequence rep-resenting each species was in the database The secondstep after OTU clustering was selecting a single represen-tative sequence from each OTU to be used in assignmentof taxonomy to each OTU For assignment of the repre-sentative sequences to sequences in the reference data-base we also used uclust at a 97 identity thresholdFollowing assignment of taxonomy to each OTU wethen output a JSON Biom table containing the taxonomicassignments of each OTU observed along with counts forthe number of sequences found in each OTU summa-rized for each of the 36 samples The Biom table wassubsequently converted to a tab delimited table withindividual counts of the plants observed in each sampleWe summarized values for each of the 12 individual

Figure 2 The rate of correct identification for increasingly largerportions of the rbcL-mini-barcode as applied to the reference se-quence data from the SCBI plant community The rate of correctidentification for a given size range of rbcL is plotted on Y-axis sizeof the tested size range of rbcL is plotted on X-axis A 20 bp slidingwindow was applied to the 540 bp rbcL reference sequences andused to test how well increasingly larger fragments of rbcL distin-guish at different taxonomic levels (family genus species) Thepoints represent the average and standard deviation of the rate ofidentification for each size range tested Also plotted are the ratesof species identification for the rbcL-mini-Barcode used in thisstudy and for the full length rbcL DNA barcode that is recognizedby CBOL



deer see Supporting InformationmdashTable S1 and thistable was used to construct Fig 3 a stacked bar-plot ofthe plant species observed for each of the 12 individualdeer where the three replicates per deer were averagedIn Supporting InformationmdashTable S1 each of the 12samples were further summarized by including only spe-cies that were present at gt1 representation withinthat sample Figures were made in either excel (Fig 4) orthe R package using ggplot

After construction of the Biom table we then alignedthe reference rbcL sequences to make a phylogenetic treethat could be used to estimate the relative genetic alphaand beta diversity among samples Alignment of referencesequences was conducted using the alignment programMUSCLE (Edgar 2004) The aligned sequences were thenapplied to build a phylogeny of the sequences withFastTree (Price et al 2009) This molecular phylogeny canbe used to describe the diversity of plants eaten in eachsample such that we can quantify the diversity of allplants found within and between fecal samples (seeErickson et al 2014 for discussion of use of rbcL barcodephylogeny for quantifying community phylogenetic diver-sity) While a molecular phylogeny for a wide diversity ofangiosperms using only a mini-barcode is imperfect therbcL locus is one of the most informative loci in recon-structing phylogenetic relationships among plant groups(Erickson et al 2014) Beta diversity defined as geneticdistance among the 36 samples was calculated in Qiimeusing the ldquoBeta_diversitypyrdquo algorithm which used theBiom table and the phylogenetic tree of reference samplesto estimate a matrix of the phylogenetic distance amongof each of the 36 samples using weighted UniFrac dis-tance estimation (Lozupone and Knight 2005)

Whole genome sequencing

Because a number of published studies have suggestedthat PCR bias can lead to skewed estimates of commu-nity composition in microbes (eg Acinas et al 2005) wesubmitted the one replicate from each of the 12 deer inaddition to a negative control for whole genomic shot-gun (WGS) sequencing using the TruSeq kit from

Figure 3 Stacked bar-graph showing the frequency of different plant species found in each of the 12 deer individuals Only the 15 most com-mon species which comprise more than 95 of all diversity are graphed For each of the 12 individual deer the three sub-replicates werecollapsed into a single value by averaging counts of the plants observed Exotic species are identified with an asterisk

Figure 4 Relative occurrence of plant species in 12 fecal samplesand in 160 1 1 m vegetation plots in the same region Thedashed line indicates the equal detection of plants in both sam-ples Points above the line indicate higher occurrence in the fecalsamples than the plant community and points below the line indi-cate lower occurrence in the samples than would be predictedfrom the plant community samples Those points representing in-vasive exotic species are identified to genus The figure is derivedfrom values presented in Supporting InformationmdashTable S2



Illumina For each of the 12 samples 300 ng of DNA wassheared using Covaris shearing matrices and the size se-lected to include a range of 300ndash600 bp fragments usingAmpure beads We used a 08 volume of Ampure beadsto remove larger fragments and a second selection us-ing 18 volume of Ampure beads to remove smallerfragments Libraries were prepared using the TruSeq kitand the resulting DNA was sequenced using a MiSeq anda V2 500 bp kit Analysis of the shotgun data followedthe same quality filtering as above using CLC BioGenomics Workbench (ver 74) but reads were notmerged and the forward and reverse sequences werekept separate We then identified five complete chloro-plast genomes that corresponded to species (only genusin the case of Quercus) observed in the PCR based assays(see above) which included (species and GenBank acces-sion Quercus aliena NC_0267901 Lonicera japonicaNC_0268391 Elaeagnus macrophylla NC_0280661Festuca arundinacea NC_0117132 and Vitis aestivalisNC_0294541) We then used CLC Bio GenomicsWorkbench (ver 74) to map the reads from each of the12 WGS sequence sets back to each of the 5 completechloroplasts using reference mapping algorithm This al-lowed us to quantify if sequence data from WGS con-tained any information from the chloroplast that wouldallow us to identify the diet of the animalsmdashwith the ex-pectation that sequence data from WGS would not sufferany of the problems that arise from PCR bias when itcomes to identifying how much of each species the ani-mals consume

Native versus exotic species consumption

We analysed the diet selection of deer in three waysFirst we compared the relative abundance of plant spe-cies observed in the diet to the relative plant abundanceas detected in an understory plant survey of the studyarea For comparison between environmental samplesand fecal DNA we combined species from the same ge-nus when there was uncertainty in species identificationat the seedling stage (eg Lonicera Acer Rubus)

Second we constructed a bipartite graph using theOTU table from Qiime and implemented the bipartitepackage in R (Dormann et al 2009) To improve legibilityof the graph only the 25 most abundant plant speciesobserved in deer diet were included in the bipartite anal-ysis (these 25 species accounted for more than 95 ofall observations) For the 12 deer each of the three repli-cates were averaged so the graph contained 12 deermapped to 25 plant species (Fig 5) The bipartite graphvisually represents the symmetry or lack thereof in dietchoice for each of the 12 individual deer We also usedthe nullttest option within the Bipartite software

package to compare the symmetry of connections withinour network to that of a randomly distributed networkwith the same number of observations using nfrac1425

A third analysis was to construct a nonmetric MDSgraph using the multidimensional scaling function in theR ldquostatsrdquo package The nMDS graph provides a visual rep-resentation of the similarity in diet of the 12 deer individ-uals and shows how the three replicates per individualmay group together The distance table producedthrough the Beta_diversitypy analysis was used as inputfor the R package We further included a random samplein the nMDS graph The random sample was generatedby partitioning the 100 000 OTU observations evenlyamong the 139 different plants that were observed inthe diet of the 12 deer Thus for the random sampleeach of the plants in the matrix was assigned 719counts as if an individual deer ate every available plantat an equal rate To evaluate whether sample origin sig-nificantly influenced total community composition weemployed non-parametric PERMANOVA analysis of simi-larity (ANOSIM) and distance based redundancy analysismodules available in Qiime (using the Vegan R package)with 1000 permutations for each test In this way wecontrasted within versus between variance in diet com-position for the 12 deer and their sub-replicates

Results

DNA isolation and PCR and sequencing

Use of the ZR Fecal DNA MiniPrepTM provided high DNAyield and successful PCR amplification with the rbcL-mini-barcode The DNA concentration of samples rangedfrom 30 to 60 ngmL and PCR using the rbcL-mini-barcodewas successful for all 36 samples using Illumina fusiontagged primers The number of sequences obtained foreach sample ranged from 105 271 to 582 045 fromwhich 100 000 sequences per sample were sampled foruse in OTU assessment following Yoccoz et al (2012)

The two microsatellite markers developed byAnderson et al (2002) were applied to each sample andno two of the 12 deer shared the exact same genotypeand thus each of the 12 individuals were unique

DNA mini-barcode performance

The mini-barcode based on the rbcL gene was highly ef-fective both in PCR success and in resolving speciesidentity in the LFDP [see Supporting InformationmdashTable S1] The rates of correct identification for speciesgenera and families (Fig 2) demonstrates that althoughthe mini-barcode marker was nearly 30 smaller thanthe full length official DNA barcode for rbcL it captured90 of the identification power of the full length DNA



barcode The full length DNA barcode was able to distin-guish among species 82 of the time in the LFDPwhereas the mini-barcode correctly assigned identity72 of the time on the same set of species The identi-fication rate is a function of the database used and werecognize that the 82 identification rate is for speciesin our SCBI reference library and does not reflect identi-fication rates for all plants The 72 identification ratereflects that 26 of the species in the reference data-base were identified to the level of genus or family butdoes not mean they will be arbitrarily assigned to an-other taxon In total all species could be identified tospecies (n frac14173) Genus (nfrac1446) or Family (nfrac1415) andthe 15 species that could only be identified to Familywere all from the Poaceae We note that reducing thethreshold for identification did not affect rates of identi-fication and the rate of correct identification at a 1 and 2 threshold were the same as a 3 thresholdFigure 2 displays the rate of correct identification forthe mini-barcode when applied to the SCBI communityand shows how smaller to larger size samples of the

rbcL barcode provide increasing rates of correct identifi-cation for different taxonomic levels The smallest sizerange of 100 bp provided accurate identifications tofamily and at 300 bp to genus The rate of identificationfor Species saturates at 80 using the full-lengthDNA barcode thus the observed rate of 72 identifica-tion at a 3 sequence divergence for the mini-barcode captures most of the diagnostic information inthe gene Our use of a 3 threshold for discriminatingamong species also means that sequencing error is un-likely to result in incorrect taxonomic assignments Wenote that the amplicon size of 379 is larger than someother mini-barcode markers notably the P6-loop em-ployed by Soininen et al (2009) and Kartzinel et al(2015) Just as more work to quantify PCR bias isneeded so too is work to establish whether there is lossof signal from more degraded diet items This will berelevant if for example there are changes in diet pref-erence over time such that older more degraded DNAreflect legitimate diet choice as animals alter their for-age preferences

Figure 5 A bipartite graph showing the foraging network for 12 individuals of deer Only the 25 most common plant samples (which capturemore than 95 of the observations) are mapped Native and exotic species are indicated by green and black colours respectively The graphis quantitative with the width of the bar connecting deer to plants corresponding to the relative proportion of that plant found in each deerrsquosdiet Invasive exotic species are identified with an asterisk



OTU diversity and structure

The deer diets had between 61 and 91 distinct OTU(meanfrac1471) although the abundance was highlyskewed to a few plant species (Fig 3) When we excludespecies detected in lt1 of the OTU our species dropsto 4ndash12 distinct OTUs per sample (meanfrac1483) The mostabundant species found in their fecal samples wereQuercus sp (we could not distinguish among the DNAsamples from the different oak species and considerthem as a single OTU) Elaeagnus umbellata Rubus phoe-nicolasius Agrostis sp Rubus flagellaris Symphoricarporbiculatus Rubus pensilvanicus Duchesnea indicaFestuca sp Poa sp and Crataegus chrysocarpa Quercussp alone accounted for 46 of all observations whensummed across all samples with no other species occur-ring at more than seven percent The 12 most abun-dantly observed species accounted for 95 of all reads[Fig 3 and see Supporting InformationmdashTable S1]Although there was a high degree of asymmetry in dietthere was a very high degree of selectivity to the fullspectrum of plants available in their environment [seeSupporting InformationmdashFig S1] In total three plantfamilies accounted for a large majority of the dietFagaceae (46 ) Poaceae (12 ) and Rosaceae (29 )even though the abundance of other families (egAsteraceae) is high in our census The negative controlproduced a weak PCR amplicon which was sequencedwith other samples but no plant sequences were ob-served and nearly all sequences were assigned to bacte-rial reference samples

Native versus exotic forage

The forest-wide inventory of plant species detected 335vascular plant species within SCBI The 1 1 m plot sur-vey detected 103 species within the vicinity of the fecalcollections [see Supporting InformationmdashTable S2] Ofthese species 29 were detected in both the fecal sam-ples and the 1 1 m plots [see SupportingInformationmdashTable S2] A comparison of plant relativeabundance in fecal and plot samples found little conver-gence [Fig 4 and see Supporting InformationmdashTableS2] Most species were either more abundant in fecal orplot samples with a significant mean difference betweenthe paired samples (X difference in sam-plesfrac14040 6 0028 SE Plt0001 nfrac14149) For thoseplant species found in at least 10 of the plot surveys orfecal samples there were 13 species that were at leasttwice as abundant in the plot surveys than the fecalsamples and 92 species of plants that were at least twiceas abundant in the fecal samples as the plot surveysMost of the latter were species commonly found in oldfields

Native plants were closer to equivalence in the twosamples than exotic plants (Xdifference in abun-dancefrac14037 and 054 respectively dffrac14147 tfrac14232Pfrac14001) but there was no pattern with respect to exoticspecies Some weregt25 more abundant in the fecesthan the plot samples (ie Lonicera sp Elaeagnus spRubus phoenicolasius Rosa multiflora) and some wereabsent or rare from fecal samples but more common inplots (gt025 ie Alliaria petiolata Lespedeza spOplismenus hirtellus Berberis Periscaria perfoliata) Wedistinguish between invasive and introduced exotic spe-cies [see Supporting InformationmdashTable S2] Counts ofthe 25 most abundant native versus exotic species fromthe species table showed that over 65 of all sequencesassigned to OTU were assigned to native species (Fig 5)However when Quercus sp were excluded half of the re-maining sequence data were assigned to non-nativespecies with 30 of OTUs being invasive exotic speciesand 20 introduced exotic species [see SupportingInformationmdashFig S1]

Bipartite network

A quantitative network mapping of deer to the 25 mostcommon plant species showed a high proportional oc-currence of Quercus in the deer diet but also demon-strated a very wide diet breadth (Fig 5) includingspecies from 12 different plant families The distributionof species found in their diet was highly asymmetricalwhen tested with nullttest in the bipartite R package(t frac142107 Plt0001)

Nonmetric MDS

The nonmetric multidimensional scaling graph (Fig 6)demonstrates the clustering of deer along the secondMDS axis relative to the randomly distributed sample re-flecting that the diet of deer is highly non-random Allthree methods used to quantify variance within and be-tween samples (PERMANOVA pseudo Ffrac14173 Pfrac140004ANOSIM r2frac14027 Pfrac140002 and DBRDA pseudo Ffrac14166Pfrac140006) found that the variance between samples washigher than variance among sub-samples We likewisedetected significant associations between communitycomposition (using the Bray-Curtis distance metric) andsample origin meaning sub-samples from the same fe-cal pile were more similar to each other than they wereto samples derived from other fecal piles As such theseresults reflect that there is greater variance among indi-vidual deer samples than there is within sub-replicatesand that a single sample from each fecal pile wassufficient to describe the diversity of diet for that animal




For all of 60 combinations of WGS reads (12 samples five chloroplast genomes) using 100 000 sequence readsper sample we found that none of the reads mapped tothe chloroplast genomes The negative control run inparallel for the WGS library failed to produce a sequencelibrary that could be sequenced and was not included inthe sequencing run In contrast to the WGS reads fromthe fecal samples we have observed in other analysesthat WGS reads from plant tissues map to chloroplastreferences at a rate of 05ndash2 (unpublished data) usinga reference mapping strategy in CLC Bio GenomicsWorkbench as we conducted here Thus the failure toidentify reads was likely attributable to plant DNA beingvery rare and the fecal DNA being dominated by hostand microbial DNA

Discussion

This is a metagenomic study of forage selection bywhite-tailed deer on exotic and native plant species Ourobservations are not conclusive but do indicate selectivefeeding by deer (Fletcher et al 2001 Keane and Crawley2002 Frerker et al 2014 Averill et al 2016) and are par-ticularly relevant given that a number of widespread ex-otic species were relatively rare in the samplesexamined Our results correspond closely to those foundby Averill et al (2016 2017) of some exotic plants suchas Alliaria petiolata and Microstegium vimineum beingunder represented in deer fecal samples and other ex-otic plants such as Lonicera japonica and Rosa multiflorabeing over represented in the same fecal samples We

could discern no common trait on which exotic plantspecies were not consumed by deer

WGS versus PCR

We tested both a PCR amplicon approach and a whole-genome-shotgun (WGS) approach to determine if PCRapproaches exhibit bias in the OTU that can be observedA number of studies have suggested that PCR can oftenlead to altered estimates of the number and abundanceof OTU in communities that are very diverse (Larsen et al2014 Wood and Salzberg 2014) Methods that use WGSto infer microbial diversity have been able to circumventthe potential bias observed in PCR (eg Rusch et al2007) In addition WGS data from fecal samples hasbeen used successfully to quantify the plant diet of a pri-mate herbivore Pygathrix nemaeus (Srivathsan et al2015) using very deep sequencing (over an order of mag-nitude greater than used in our study) However our re-sults found that a PCR amplicon strategy revealed a widediversity of plant species whereas the WGS approachwas unable to recover and assign sequence data towhole genome reference for species that were com-monly observed in the PCR amplicon analysis When wethen mapped WGS reads from each of the 12 deer indi-viduals to the five complete chloroplast genomes usingboth MIRA and CLC Bio Genomics Workbench we foundno sequences could be assigned to the plant chloro-plasts and that sequences were overwhelmingly micro-bial These results strongly contrast with Srivathsan et al(2016) and may be attributable to our focal herbivorespecies being a ruminant versus a primate in addition todifferences in the number of WGS sequences used Wenote that Covaris shearing of the fecal DNAs may havefurther degraded the plant sequences which were likelysmall to begin with making them too small to be recov-ered in the library processing steps after shearingAlternatively we observed a wide number of plant spe-cies using PCR amplicon methods [Fig 3 and seeSupporting InformationmdashTable S1] The rbcL mini-barcode recovered sequence data throughout theflowering plant clade from ferns to Asterids and Rosidssuggesting little bias in recovery using that marker Thegene did have limits in assigning species designationto the grass family (Poaceae) a majority of the OTUsthat could not be assigned to species were from thisfamily

Marker selection and identification rate

We employed a rbcL based mini-barcode PCR marker de-rived from a subset of the larger rbcL DNA barcode locus(CBOL Plant Working Group 2009) This mini-barcodecaptured a majority of the identification power (nearly

Figure 6 A nonmetric multidimensional scaling (nMDS) graph ofthe weighted UniFrac distances among the 36 deer samples isplotted Each of the 36 points is labelled and a reference samplecorresponding to completely random diet (see text) from the LFDPcommunity is plotted for reference The three sub-samples fromeach of the 12 individual deer are the same colour



90 ) of the complete DNA barcode (Fig 2) but with70 the size of the amplicon Smaller amplicon size isknown to improve recovery of increasingly degradedplant samples (Valentini et al 2009) such as what maybe found in scat used to deduce herbivore diet The rateof identification within the SCBI plot for the mini-barcodewas 72 which is very similar to the overall rate of iden-tification reported for the combined two locus plant DNAbarcode advocated by CBOL (76 resolution)(CBOLPlant Working Group 2009) The rate of identification isalso similar to the P6-loop advocated by Taberlet andcolleagues (Valentini et al 2009 Pompanon et al 2012)However the much larger available database of rbcL se-quence data due to its assignment as the official DNAbarcode locus and its historical use in plant molecularsystematics means that many more species have beensequenced at this locus and may be more likely uniquelyidentified using rbcL We also observed that amplifica-tion of the rbcL mini-barcode was not affected by addi-tion of a phusion primer adaptor facilitating NGS librarypreparation Lastly although the P6 marker has beensuccessfully used where larger PCR amplicons fail (egKartzinel et al 2015) the marker is highly A-T rich andcontains many short mononucleotide repeats that mayaffect sequencing error rates and hence correct taxo-nomic assignment We believe that use of the rbcL mini-barcode reported here will best serve future studies thatwish to diagnose plant-herbivore trophic interactionsWe did not perform explicit in-silico testing of primer biaswith our rbcL mini-barcode and we note that such anal-ysis is appropriate for subsequent analyses Published re-ports that describe primer bias have focused on loss ofrelatively rare and disjunct species and as such wouldnot likely affect our conclusions In addition primermatching to common invasive species that were notcommon in our diet analysis such as Alliaria petiolataVitis rotundifolia and Rosa multiflora showed that theprimers matched them perfectly and thus their preva-lence in our diet observations was likely not due to PCRamplification bias

There were some problematic assignments from theDNA process These fall into two groups the first areplants commonly associated with old fields that werefound in fecal samples but not detected in our forest plotsamples This group includes several of the grasses bothnative and exotic The second instance is for easternhemlock (Tsuga canadensis) which is a native speciesdetected in two fecal samples but is not common in theareas and the closest known individuals would bearound residential houses within a 2 km of the collectionsites The field sampling was conducted at the height ofthe acorn fall and a previous study at the site recorded

deer moving up to 3 km to consume acorns within thisstand (McShea and Schwede 1993)

Sampling in the field

The foraging behaviour of the animals and the relativelysmall volume of sample taken from each fecal pelletgroup prompted questions that it would be difficult tofully characterize the diet breadth of an animal from asingle small subsample To test this concern we subdi-vided each fecal sample into three sub-replicates andthen sequenced each independently This produced 36samples from our original 12 samples We further soughtto minimize PCR bias through repeating the PCR twice foreach of the 36 samples and then pooling each of the PCRreplicates in equal concentration We then examinedwithin and between sample variance for the differentsamples The direct analysis of the variance within andbetween samples pointed toward greater varianceamong individuals compared to variance within the threereplicates taken from each individualrsquos fecal sample Asnew species were still being detected in the third sub-sample of several fecal samples we cannot closely esti-mate the number of subsamples needed to exhaustivelycharacterize diet but the higher variance between sam-ples than within indicates we are approaching the maxi-mum needed The MDS projection which characterizesthe diversity of the entire set of plants found in eachsample similarly pointed to the clustering of the deerdiets relative to a random sample suggesting that whiledeer diets are diverse they are not random and thus asampling strategy using a single sub-sample per fecal issufficient A larger unknown is the length of time planttissues reside within the gut we speculate that the verylarge diet breadth observed reflects a significant time(gt1 days) as well as reflecting the sensitivity of themethods

We discount the role that pollen may have played inincreasing the observed number of OTU as collectionswere made in the fall when relatively few species are inflower However we acknowledge that as with any typeof environmental sequencing that our samples may in-clude DNA from pollen soil or other natural sources Ourobservations suggest a much higher number of plants inthe diet compared to reports from microhistologicalstudies (Kowalczyk et al 2011) For a highly mobile gen-eralist herbivore the sensitivity of the metagenomicmethods likely reveal legitimate patterns of widespreadselective browsing

Low frequency OTU

We observed a very high number of OTU in our study al-though their distribution was highly asymmetric The 15



most common species accounted for 95 of all OTUcounts and the remaining observations recorded rela-tively few counts and their observation did not signifi-cantly affect our quantification of native versus exoticplant forage choice However the issue of low frequencyOTU is important and reflects more generally how weidentified true versus false observations A number ofstudies have explored this issue (de Barba et al 2014Ficetola et al 2016) and in cases where there is not a highlevel of asymmetry their sampling methodology may en-sure more stringent OTU assignment Those studies cor-rectly note that caution should be taken where single readcounts are used as data and closely related species areconsidered in the reference database In these cases mu-tations introduced through PCR or sequence error cannotbe distinguished from legitimate low frequency OTU ob-servations or even background environmental contamina-tion For our purposes the high number of low frequencyOTU likely reflect a combination of all of these processesWhen we exclude species whose OTU units are lt1 oftotal units for each sample we do greatly reduce thenumber of species detected However we note that mostspecies we identified at SCBI exceeded the 3 thresholdfor sequence differences and that negative controls usedin both the PCR and WGS assays failed to contain plant se-quence data suggesting that laboratory contaminationwas not a factor

Quantifying trophic interactions

The critical role of herbivory in structuring natural com-munities is widely accepted (Janzen 1970 Huntly 1991Danell et al 2006) although herbivory is implicated inboth promoting the diversity of plant communities(Janzen 1970 Crawley 1983 Bowers 1993 Parks et al2005) and reducing diversity through promoting thespread of invasive species due to herbivore preference ofnative species (Keane and Crawley 2002 Jogesh et al2008) Experimental trials of animals in captivity or ex-clusion experiments have found evidence for both in-creasing and decreasing diversity in naturalcommunities (Augustine and McNaughton 1998 Coteet al 2014 Kalisz et al 2014) which suggests that her-bivory may indeed be inconsistent in its effects

Our observation suggests that white-tailed deer areselective feeders (Figs 3ndash5) The effect of selecting nativespecies in higher proportion than they exist in the fieldwas not observed for all native plants and the dietchoice of the deer was clearly focused on a few speciesbut reflected preference for native species The wide-spread success of invasive plants in the local floristiccommunity particularly Alliaria petiolata may well be

promoted in part due to the selective browsing of nativespecies by white-tailed deer

Along with competitive advantages for light and wa-ter and altered disturbance regimes selective browsingof herbivores is a primary agent for successful invasionsof exotic species (Ricciardi et al 2013) We report evi-dence for selective browsing by deer with 105 commonspecies being detected in feces at least 25 more orless than found in plot surveys The list of these selectedspecies includes several ecologically important woodyspecies such as Quercus Ulmus Acer Juglans andViburnum (McShea et al 2007) Given that deer are highlygeneralist feeders they are likely less affected bychanges in community composition but our increasingability to measure and quantify diet from non-destructive methods such as through fecal DNA will al-low us to better quantify how changes in plant commu-nity structure may affect network dynamics in naturalcommunities

Conclusions and Future Challenges

In conclusion we observed a wide diversity of plant spe-cies in the diet of 12 individuals of white-tailed deer col-lected at the SCBI long term research site [Fig 3 and seeSupporting InformationmdashTable S1] Fecal pellets hadtotal genomic DNA extracted using the Zymo fecal DNAkit with the resulting DNA PCR amplified using a rbcL-mini-barcode marker set first described in this paperThe rbcL mini-barcode marker exhibited a high rate ofspecies level identification and could resolve all se-quences from plants found at the site to at least genus ifnot species (Fig 2) The availability of an exhaustive ref-erence library of rbcL sequences compiled for the SCBIsite allowed us to use reference based assignment of re-covered rbcL sequences and demonstrated a wide rangeof species in the diet suggesting the universality of themarkers The ability to readily recover plant DNA datafrom fecal pellets of a ruminant suggests we will be ableto directly quantify plant diets and address hypothesesregarding the influence of ungulate herbivores on thestructure of natural plant communities

The use of metagenomics to assess diet of ungulateshas great promise as it allows for species-specific identi-fication of forage plants in natural communities Thetechniques use by the research community will dependon several factors Our ability to identify 72 of OTUs tospecies may be sufficient to determine native or exoticspecies consumed by deer but may not satisfy require-ments for studies of grazing species where the bulk ofthe diet is rapidly evolving grass species in that case useof a second locus barcode may be necessary The



reagent costs are not high ($20sample) but the exper-tise and access to sequencing equipment will play alarger role in determining the final cost Its adoption alsorelies on an established reference library which we hadfor this study site and the rapid collection of fecal sam-ples prior to degradation which was facilitated by thefield fecal DNA kits used In addition there still remainsome major technical questions before broadening theuse of this technique We have assumed that theamount of OTUs found in the feces reflects the propor-tion of the plant in the diet Studies using coarse samplesand visual inspection of prepared slides have not foundthis assumption to be valid However the sensitivity ofthe metagenomics method infers that we are not heavilyreliant on examining only large undigested plant seg-ments This technique should be far more reliable at de-tecting plant consumption but those tests remain to beconducted We are not sure of the amount of samplingneeded to characterize the single fecal pile of an individ-ual or the daily diet We collected three samples for eachfecal pile and these samples showed remarkable consis-tency but as could be expected were not in agreementfor rare OTUs A larger question was the forage period asingle fecal pile represents we found a high diversity ofplant species in single fecal piles but are not sure if thesesamples represent an hour day week or month of forag-ing With a complex digestive track and regurgitation ofplant materials to increase digestion we do not knowwhat time period we are sampling for this ungulate Acontrolled experiment with captive deer placed on spe-cific diets would answer these questions

Sources of Funding

David L Erickson and Padmini Ramachandran werefunded as an ORISE Fellow by the Office of RegulatoryScience (ORS) Centre for Food Safety and AppliedNutrition (CFSAN) US Food and Drug AdministrationFunding for DNA sequencing was provided by theDivision of Microbiology ORS CFSAN US Food and DrugAdministration Funding for DNA Sequencing was pro-vided through support to the Division of Microbiology atthe Food and Drug Administration The plant DNABarcode Library and the plant inventory were completedthrough the Smithsonian Institutionrsquos Global GenomicsInitiative via a Smithsonian DNA Barcoding Networkgrant

Contributions by the Authors

The experiment was designed by David Erickson andWilliam McShea The field data was collected by William

McShea and Norman Bourg DNA sequencing and inter-pretation was completed by David Erickson PadminiRamachandran and Andrea Ottesen Data Analysis in-cluded William McShea David Erickson and ElizabethReed The manuscript was prepared by William McSheaDavid Erickson Norman Bourg and Andrea Ottesen DNAsequence data are available from DL Erickson on requestSupporting InformationmdashTable S1 is a summary ofthese data

Acknowledgements

Additional field and data assistance was provided byXiaoli Shen Chad Stewart Olivia Cosby and BrandonRoss

Conflict of Interest Statement

All authors had no conflict of interests Any use of tradefirm or product names is for descriptive purposes onlyand does not imply endorsement by the USGovernment

Supporting Information

The following additional information is available in theonline version of this articlemdash

Figure S1 The percent of total OTUs for the 25 mostcommon plant species for each of 36 fecal samples (3samples from 12 fecal piles) Sample ID indicates the fe-cal pile and the subsample (eg Sample 11 12 13 are 3sub-samples of the same fecal pile) These 25 speciesrepresentgt95 of the plant species detected in thesamples Exotic species are identified with an asterisk

Table S1 List of species identified based on compari-son of DNA samples extracted from fecal pellets from 12deer when compared to reference DNA barcode createdfor site For each of the 12 deer 3 sub-samples were pre-pared from each fecal pile and the sum of OTUs for eachspecies listed in table For those deer where the specieswas not found in each sub-sample we have indicated iftwo () or one () sub-samples contained the plant spe-cies Some OTUs could not be classified to species leveland only the genus level is indicated Species are listed intheir taxonomic order The bottom of each column sumsthe total number of species detected and the number ofspecies if values less than 1 of the total OTUs areexcluded

Table S2 List of species detected and their relativeabundance in either 160 1 1 m plots within the forestaround the Large Tree Plot or within the 12 fecal samplesobtained in the same general location When specieslevel identification was not possible for either the fecal



sample or the individual plant we have listed the genusPlants are listed based on their relative abundance in theplot survey We have categorized plants as native intro-duced or invasive The data in this table are used to cre-ate Fig 4 in Paper

Literature CitedAcinas SG Sarma-Rupavtarm R Klepac-Ceraj V Polz MF 2005 PCR-

induced sequence artifacts and bias insights from comparisonof two 16S rRNA clone libraries constructed from the same sam-ple Applied and Environmental Microbiology 718966ndash9

Alipayo D Valdez R Holechek JL Cardenas M 1992 Evaluation ofmicrohistological analysis for determining ruminant diet bo-tanical composition Journal of Range Management 1148ndash152

Anderson JD Honeycutt RL Gonzales RA Gee KL Skow LCGallagher RL Honeycutt DA DeYoung RW 2002 Developmentof microsatellite DNA markers for the automated genetic char-acterization of white-tailed deer populations The Journal ofWildlife Management 6667ndash74

Augustine DJ McNaughton SJ 1998 Ungulate effects on the func-tional species composition of plant communities herbivore se-lectivity and plant tolerance The Journal of WildlifeManagement 621165ndash1183

Averill KM Mortensen DA Smithwick EA Post E 2016 Deer feedingselectivity for invasive plants Biological Invasions 181247ndash1263

Averill KM Mortensen DM Smithwick EAH Susan Kalisz S McSheaWJ Bourg NA Parker JD Royo AA Abrams MD Anurag AAgrawal AA Apsley DK Blossey B Caraher KL Johnson SEMasson R Nuzzo VA Snitzer J 2017 Native deer browsing in-creases the degree of plant invasion AoBPlants

Baker BJ Banfield JF 2003 Microbial communities in acid minedrainage FEMS Microbiology Ecology 44139ndash152

Blaxter M Mann J Chapman T Thomas F Whitton C Floyd R AbebeE 2005 Defining operational taxonomic units using DNA bar-code data Philosophical Transactions of the Royal Society BBiological Sciences 360 1935ndash1943

Blaser MJ 2014 The microbiome revolution The Journal of ClinicalInvestigation 1244162ndash4165

Bourlat SJ Borja A Gilbert J Taylor MI Davies N Weisberg SBGriffith JF Lettieri T Field D Benzie J Glockner FO 2013Genomics in marine monitoring new opportunities for assessingmarine health status Marine Pollution Bulletin 7419ndash31

Bourg NA McShea WJ Thompson JR McGarvey JC Shen X 2013Initial census woody seedling seed rain and stand structuredata for the SCBI SIGEO Large Forest Dynamics Plot (EcologicalArchives EO94-195) Ecology 942111

Bowers MA 1993 Influence of herbivorous mammals on an old-field plant community years 1ndash4 after disturbance Oikos 1129ndash141

Caporaso JG Kuczynski J Stombaugh J Bittinger K Bushman FDCostello EK Fierer N Pena AG Goodrich JK Gordon JI HuttleyGA 2010 QIIME allows analysis of high-throughput communitysequencing data Nature Methods 7335ndash336

CBOL Plant Working Group Hollingsworth PM Forrest LL Spouge JLHajibabaei M Ratnasingham S van der Bank M Chase MWCowan RS Erickson DL Fazekas AJ 2009 A DNA barcode for

land plants Proceedings of the National Academy of SciencesUSA 10612794ndash12797

Clare EL 2014 Molecular detection of trophic interactions emerg-ing trends distinct advantages significant considerations andconservation applications Evolutionary Applications 71144ndash1157

Cote SD Beguin J de Bellefeuille S Champagne E Thiffault NTremblay JP 2014 Structuring effects of deer in boreal forestecosystems Advances in Ecology doi1011552014917834

Cote SD Rooney TP Tremblay JP Dussault C Waller DM 2004Ecological impacts of deer overabundance Annual Review ofEcology Evolution and Systematics 31113ndash147

Crawley MJ 1983 Herbivory the dynamics of animal-plant interac-tions Oxford Blackwell Scientific Publications

Creed RP 2000 Is there a new keystone species in North Americanlakes and rivers Oikos 91405ndash408

Cumplido JH Glauser G Benrey B 2016 Cascading effects of early-season herbivory on late-season herbivores and their parasit-oids Ecology 971283ndash1297

Danell K Bergstrom R Duncan P Pastor J 2006 Large herbivoreecology ecosystem dynamics and conservation CambridgeCambridge University Press

Deagle BE Kirkwood R Jarman SN 2009 Analysis of Australian furseal diet by pyrosequencing prey DNA in faeces MolecularEcology 182022ndash2038

Deagle BE Chiaradia A McInnes J Jarman SN 2010 Pyrosequencingfaecal DNA to determine diet of little penguins is what goes inwhat comes out Conservation Genetics 112039ndash2048

de Barba M Miquel C Boyer F Mercier C Rioux D Coissac E TaberletP 2014 DNA metabarcoding multiplexing and validation ofdata accuracy for diet assessment application to omnivorousdiet Molecular Ecology Resources 14306ndash332

Dormann CF Fruend J Bluethgen N Gruber B 2009 Indices graphsand null models analyzing bipartite ecological networks TheOpen Ecology Journal 27ndash24

Edgar R 2004 MUSCLE multiple sequence alignment with high ac-curacy and high throughput Nucleic Acids Research 321792ndash1797

Enge S Nylund GM Harder T Pavia H 2012 An exotic chemicalweapon explains low herbivore damage in an invasive algaEcology 932736ndash2745

Erickson DL Jones FA Swenson NG Pei N Bourg NA Chen W DaviesSJ Ge XJ Hao Z Howe RW Huang CL 2014 Comparative evolu-tionary diversity and phylogenetic structure across multiple for-est dynamics plots a mega-phylogeny approach Frontiers inGenetics 5358

Ficetola GF Taberlet P Coissac E 2016 How to limit false positivesin environmental DNA and metabarcoding Molecular EcologyResources 16604ndash607

Fletcher JD Shipley LA McShea WJ Shumway DL 2001 Wildlifeherbivory and rare plants the effects of white-tailed deer ro-dents and insects on growth and survival of Turkrsquos cap lilyBiological Conservation 101229ndash238

Frerker K Sabo A Waller D 2014 Long-term regional shifts in plantcommunity composition are largely explained by local deer im-pact experiments PLoS One 9e115843

Garcıa-Robledo C Erickson DL Staines CL Erwin TL Kress WJ 2013Tropical plant-herbivore networks reconstructing species inter-actions using DNA barcodes PLoS One 8e52967



Goldberg Caren S Strickler KM Pilliod DS 2015 Moving environmen-tal DNA methods from concept to practice for monitoringaquatic macroorganisms Biological Conservation 1831ndash3

Hajishengallis G Darveau RP Curtis MA 2012 The keystone-pathogen hypothesis Nature Reviews Microbiology 10717ndash725

Hibert F Taberlet P Chave J Scotti-Saintagne C Sabatier D Richard-Hansen C 2013 Unveiling the diet of elusive rainforest herbi-vores in Next Generation Sequencing era The tapir as a casestudy PLoS One 8e60799

Hugenholtz P Goebel BM Pace NR 1998 Impact of culture-independent studies on the emerging phylogenetic view of bac-terial diversity Journal of Bacteriology 1804765ndash4774

Holechek JL Vavra M Pieper RD 1982 Botanical composition deter-mination of range herbivore diets a review Journal of RangeManagement 1309ndash315

Holm JA Thompson JR McShea WJ Bourg NA 2013 Interactive ef-fects of chronic deer browsing and canopy gap disturbance onforest successional dynamics Ecosphere 4 144

Huntly N 1991 Herbivores and the dynamics of communities andecosystems Annual Review of Ecology and Systematics 22477ndash503

Illumina Corporation 2013 16S Metagenomic Sequencing LibraryPreparation httpswwwilluminacomcontentdamillumina-supportdocumentsdocumentationchemistry_documentation16s16s-metagenomic-library-prep-guide-15044223-bpdf

Janzen DH 1970 Herbivores and the number of tree species in trop-ical forests American Naturalist 104501ndash528

Jogesh T Carpenter D Cappuccino N 2008 Herbivory on invasiveexotic plants and their non-invasive relatives BiologicalInvasions 10797ndash804

Kalisz S Spigler RB Horvitz CC 2014 In a long-term experimentaldemography study excluding ungulates reversed invaderrsquos ex-plosive population growth rate and restored nativesProceedings of the National Academy of Sciences USA 1114501ndash4506

Kartzinel TR Chen PA Coverdale TC Erickson DL Kress WJ KuzminaML Rubenstein DI Wang W Pringle RM 2015 DNA metabarcod-ing illuminates dietary niche partitioning by African large herbi-vores Proceedings of the National Academy of Sciences USA1128019ndash8024

Keane RM Crawley MJ 2002 Exotic plant invasions and the enemyrelease hypothesis Trends in Ecology and Evolution 17164ndash170

Kowalczyk R Taberlet P Coissac E Valentini A Miquel C KaminskiT Wojcik JM 2011 Influence of management practices onlarge herbivore diet case of European bison in Białowie _zaPrimeval Forest (Poland) Forest Ecology and Management 261821ndash828

Kress WJ Erickson DL 2007 A two-locus global DNA barcode forland plants the coding rbcL gene complements the non-codingtrnH-psbA spacer region PLoS One 2e508

Kress WJ Erickson DL Jones FA Swenson NG Perez R Sanjur OBermingham E 2009 Plant DNA barcodes and a communityphylogeny of a tropical forest dynamics plot in PanamaProceedings of the National Academy of Science of the USA 10618621ndash18626

Lane DJ Pace B Olsen GJ Stahl DA Sogin ML Pace NR 1985 Rapiddetermination of 16S ribosomal RNA sequences for phyloge-netic analyses Proceedings of the National Academy of SciencesUSA 826955ndash6959

Larsen MV Cosentino S Lukjancenko O Saputra D Rasmussen SHasman H Sicheritz-Ponten T Aarestrup FM Ussery DW LundO 2014 Benchmarking of methods for genomic taxonomyJournal of Clinical Microbiology 521529ndash1539

Leray M Yang JY Meyer CP Mills SC Agudelo N Ranwez V BoehmJT Machida RJ 2013 A new versatile primer set targeting ashort fragment of the mitochondrial COI region for metabarcod-ing metazoan diversity application for characterizing coral reeffish gut contents Frontiers in Zoology 101

Leray M Knowlton N 2015 DNA barcoding and metabarcoding ofstandardized samples reveal patterns of marine benthic diver-sity Proceedings of the National Academy of Sciences 1122076ndash2081

Leray M Meyer CP Mills SC 2015 Metabarcoding dietary analysis ofcoral dwelling predatory fish demonstrates the minor contribu-tion of coral mutualists to their highly partitioned generalistdiet PeerJ epub

Letnic M Koch F Gordon C Crowther MS Dickman CR 2009Keystone effects of an alien top-predator stem extinctions ofnative mammals Proceedings of the Royal Society of London BBiological Sciences 2763249ndash3256

Lozupone C Knight R 2005 UniFrac a new phylogenetic methodfor comparing microbial communities Applied andEnvironmental Microbiology 718228ndash8235

Macel M Vos RC Jansen JJ Putten WH Dam NM 2014 Novel chem-istry of invasive plants exotic species have more unique metab-olomic profiles than native congeners Ecology and Evolution 42777ndash2786

Maron JL Crone E 2006 Herbivory effects on plant abundance dis-tribution and population growth Proceedings of the RoyalSociety of London B Biological Sciences 2732575ndash2584

McShea WJ 2012 Ecology and management of white-tailed deer ina changing world Annals of the New York Academy of Sciences124945ndash56

McShea WJ Healy WM Devers P Fearer T Koch FH Stauffer DWaldon J 2007 Forestry matters decline of oaks will impactwildlife in hardwood forests The Journal of WildlifeManagement 711717ndash1728

McShea WJ Schwede G 1993 Variable acorn crops responses ofwhite-tailed deer and other mast consumers Journal ofMammalogy 74999ndash1006

McCabe TR McCabe RE 1997 Recounting whitetails past InMcShea WJ Underwood HB Rappole JH eds The science ofoverabundance Deer ecology and population managementWashington DC Smithsonian Press 11ndash26

Meier R Shiyang K Vaidya G Ng PK 2006 DNA barcoding and tax-onomy in Diptera a tale of high intraspecific variability and lowidentification success Systematic Biology 55715ndash728

Murray DC Bunce M Cannell BL Oliver R Houston J White NEBarrero RA Bellgard MI Haile J 2011 DNA-based faecal dietaryanalysis a comparison of qPCR and high throughput sequencingapproaches PLoS One 6e25776

Paine RT 1966 Food web complexity and species diversityAmerican Naturalist 165ndash75

Parker JD Burkepile DE Hay ME 2006 Opposing effects of native andexotic herbivores on plant invasions Science 3111459ndash1461

Parks CG Radosevich SR Endress BA Naylor BJ Anzinger D Rew LJMaxwell BD Dwire KA 2005 Natural and land-use history of theNorthwest mountain ecoregions (USA) in relation to patterns of



plant invasions Perspectives in Plant Ecology Evolution andSystematics 7137ndash158

Pastor J Naiman RJ 1992 Selective foraging and ecosystem pro-cesses in boreal forests American Naturalist 139690ndash705

Pereyra PJ 2016 Revisiting the use of the invasive species conceptan empirical approach Austral Ecology 41519ndash528

Pompanon F Deagle BE Symondson WO Brown DS Jarman SNTaberlet P 2012 Who is eating what diet assessment usingnext generation sequencing Molecular Ecology 211931ndash1950

Price MN Dehal PS Arkin AP 2009 FastTree computing large mini-mum evolution trees with profiles instead of a distance matrixMolecular Biology and Evolution 261641ndash1650

Richardson DM Pysek P Rejmanek M Barbour MG Panetta FD WestCJ 2000 Naturalization and invasion of alien plants conceptsand definitions Diversity and Distributions 693ndash107

Ricciardi A Hoopes MF Marchetti MP Lockwood JL 2013 Progresstoward understanding the ecological impacts of nonnative spe-cies Ecological Monographs 83263ndash282

Rooney TP 2001 Deer impacts on forest ecosystems a NorthAmerican perspective Forestry 74201ndash208

Rusch DB Halpern AL Sutton G Heidelberg KB Williamson S YoosephS Wu D Eisen JA Hoffman JM Remington K Beeson K 2007 TheSorcerer II global ocean sampling expedition northwest Atlanticthrough eastern tropical Pacific PLoS Biology 5(3)e77

Sanders NJ Gotelli NJ Heller NE Gordon DM 2003 Community dis-assembly by an invasive species Proceedings of the NationalAcademy of Sciences USA 1002474ndash2477

Shears NT Babcock RC 2003 Continuing trophic cascade effects af-ter 25 years of no-take marine reserve protection MarineEcology Progress Series 2461ndash6

Shen X Bourg NA McShea WJ Turner BL 2016 Long-term effects ofwhite-tailed deer exclusion on the invasion of exotic plants a casestudy in a Mid-Atlantic temperate forest PLoS One 11e0151825

Srivathsan A Sha J Vogler AP Meier R 2015 Comparing the effec-tiveness of metagenomics and metabarcoding for diet analysisof a leaf-feeding monkey (Pygathrix nemaeus) MolecularEcology Resources 15250ndash261

Sogin ML Morrison HG Huber JA Mark Welch D Huse SM Neal PRArrieta JM Herndl GJ 2006 Microbial diversity in the deep sea

and the underexplored rare biosphere Proceedings of theNational Academy of Sciences USA 10312115ndash12120

Soininen EM Valentini A Coissac E Miquel C Gielly L Brochmann CBrysting AK Soslashnsteboslash JH Ims RA Yoccoz NG Taberlet P 2009Analysing diet of small herbivores the efficiency of DNA barcod-ing coupled with high-throughput pyrosequencing for decipher-ing the composition of complex plant mixtures Frontiers inZoology 616

Taberlet P Coissac E Hajibabaei M Rieseberg LH 2012Environmental DNA Molecular Ecology 211789ndash1793

Thomsen PF Willerslev E 2015 Environmental DNAndashan emergingtool in conservation for monitoring past and present biodiver-sity Biological Conservation 1834ndash18

Turnbaugh PJ Hamady M Yatsunenko T Cantarel BL Duncan A LeyRE Sogin ML Jones WJ Roe BA Affourtit JP Egholm MHenrissat B Heath AC Knight R Gordon JI 2009 Acore gut microbiome in obese and lean twins Nature 457480ndash484

United States Department of Agriculture Natural ResourcesConservation Service Plants Database Retrieved from httpsplantsusdagovjava (6 Feb 2017)

Valentini A Pompanon F Taberlet P 2009 DNA barcoding for ecolo-gists Trends in Ecology and Evolution 24110ndash117

Valery L Herve F Lefeuvre J Simberloff D 2008 In search of a realdefinition of the biological invasion phenomenon itselfBiological Invasions 101345ndash1351

Vavra M Parks CG Wisdom MJ 2007 Biodiversity exotic plant spe-cies and herbivory the good the bad and the ungulate ForestEcology and Management 24666ndash72

Virginia Department of Conservation and Recreation 2015 VirginiaInvasive Plant Species List Retrieved from httpwwwdcrvirginiagovnatural-heritageinvsppdflist (6 Feb 2017)

Wood DE Salzberg SL 2014 Kraken ultrafast metagenomic se-quence classification using exact alignments Genome Biology15R46

Yoccoz NG Brathen KA Gielly L Haile J Edwards ME Goslar T VonStedingk H Brysting AK Coissac E Pompanon F Soslashnsteboslash JH2012 DNA from soil mirrors plant taxonomic and growth formdiversity Molecular Ecology 213647ndash3655



Introduction

Trophic interactions among species are thought to be acritical regulatory mechanism affecting species fitnessand community composition (Paine 1966 Shears andBabcock 2003) These interactions can take many formsthe most familiar of which include predator-prey (Letnicet al 2009) hostndashpathogen (Hajishengallis et al 2012)and plantndashherbivore (Creed 2000) These interactions ap-pear to be part of a broader cascade of interaction ef-fects all with significant impacts on individual speciesand the communities they inhabit (Cumplido et al 2016)While invertebrate herbivory has been widely studiedand implicated in structuring plant communities (Janzen1970 Maron and Crone 2006 McShea et al 2007) largeherbivores can also shift plant community compositionin both forests (Frerker et al 2014) and grasslands(Augustine and McNaughton 1998) although their role isless well understood Improved quantification of ungu-late diet will greatly contribute to our understanding ofhow foraging behaviour sustains or degrades naturalcommunities and importantly how the removal of spe-cies from communities through extinction and extirpa-tion may have a cascading effect on naturalcommunities

Increasingly molecular genetic tools to quantify tro-phic interactions have been explored from a technicalperspective (Hibert et al 2013 Ficetola et al 2016)These studies follow the pathway of metagenomic anal-yses first pioneered in environmental and human micro-biome studies of bacteria (Lane et al 1985 Hugenholtzet al 1998) using PCR amplicon data from DNA samplescontaining a mixture of species and large referencedatabases for those genes notably 16S in microbialcommunities (Baker and Banfield 2003 Sogin et al 2006Turnbaugh et al 2009) In animal studies the CO1 genethe official DNA barcode in animals has been utilized(Deagle et al 2009 Leray et al 2013) while in herbivoresdifferent markers have been used (Pompanon et al2012 Garcıa-Robledo et al 2013) Several program pack-ages have been developed to help streamline the pro-cess of taxonomic identification of the sequences inmixed samples (such as fecal samples containing plantDNA) notably Qiime (Caporaso et al 2010) Most studiesin eukaryotes have focused on accurately measuring di-versity (Leray and Knowlton 2015 Goldberg et al 2015)and have begun to quantify interactions between species(Murray et al 2011) There have been comparativelyfewer studies that directly engage in hypothesis testing(Leray et al 2015 Kartzinel et al 2015 Soininen et al2009) compared to bacterial metagenomics (Blaser2014 Thomsen and Willerslev 2015) But numerous re-views have highlighted the utility of these methods in

ecology (Valentini et al 2009 Bourlat et al 2013 Clare2014) and increasing numbers of reports implementingmetagenomic methods are being published

Studies that use metagenomics to diagnose interac-tions between different trophic levels in natural commu-nities are rare (eg Kartzinel et al 2015 Murray et al2011 Soininen et al 2009) particularly compared to thewidespread use of metagenomics to quantify organismaldiversity in an array of environments (Taberlet et al2012) Nonetheless using metagenomics in trophic in-teractions has been well reviewed (Valentini et al 2009Bourlat et al 2013 Clare 2014) Most metagenomic stud-ies of inter-trophic interactions have focused on predatordiets (eg Deagle et al 2009 2010 Bourlat et al 2013Leray et al 2013 2015) with few focusing on herbivorediet (Soininen et al 2009 Kartzinel et al 2015) Yet thepower that metagenomics offers in revealing thestrength of trophic interactions suggests that this will be-come an increasingly common tool particularly forquantifying animal diet when the animalsrsquo feedingbehaviour is not directly observable Deducing herbivorediet is relatively simple as the DNA of the herbivore itselfwill not confound the bioinformatic analysis of the plantsthey consume This is in contrast to carnivore diets wherethe DNA of the carnivore may be PCR amplified alongwith the diet of its prey (Leray et al 2013) Yet metage-nomic studies investigating trophic interactions amongherbivores and their plant community have also beenrare especially relative to the role that herbivory plays inshaping natural plant communities (Cote et al 20042014 McShea 2012 Holm et al 2013 Frerker et al2014)

Historically observations of herbivore diet have largelytaken three forms direct observation (Holechek et al1982) exclusion experiments (Bowers 1993 Parker et al2006) and micro-histological analysis of animal scats(Alipayo et al 1992 Kowalczyk et al 2011) All of thesemethods have various limitations but the use of geneticsto deduce animal diet from feces was challenging be-cause the traditional sequencing technology could onlydeduce one species at a time with accuracy thereby be-ing limited to use only with highly specialized herbivores(Garcıa-Robledo et al 2013) However for generalist her-bivores with broad diets the recent advances in mas-sively parallel Next-Generation Sequencing methodsenables us to simultaneously sequence DNA from allplant species in a mixture (Soininen et al 2009)

In the forests of North America exotic insects andplants have caused significant ecosystem alteration anda body of theory has developed around the mechanismsby which some exotic species become invasive (Sanderset al 2003) There are multiple hypotheses for invasivespecies success that rely on preferential consumption of







Methods













PCR






































Results


















Bipartite network


Nonmetric MDS






Discussion



WGS versus PCR













Low frequency OTU
















Sources of Funding





Acknowledgements




















































































































Methods













PCR






































Results


















Bipartite network


Nonmetric MDS






Discussion



WGS versus PCR













Low frequency OTU
















Sources of Funding





Acknowledgements























































































































PCR






































Results


















Bipartite network


Nonmetric MDS






Discussion



WGS versus PCR













Low frequency OTU
















Sources of Funding





Acknowledgements
















































































































































Results


















Bipartite network


Nonmetric MDS






Discussion



WGS versus PCR













Low frequency OTU
















Sources of Funding





Acknowledgements


























































































































Bipartite network


Nonmetric MDS






Discussion



WGS versus PCR













Low frequency OTU
















Sources of Funding





Acknowledgements


















































































































Discussion



WGS versus PCR













Low frequency OTU
















Sources of Funding





Acknowledgements






















































































































Low frequency OTU
















Sources of Funding





Acknowledgements




























































































































Sources of Funding





Acknowledgements
















































































































Reconstructing a herbivore’s diet using a novel rbcL DNA ... · they consume. This is in contrast...

Documents

Transcript of Reconstructing a herbivore’s diet using a novel rbcL DNA ... · they consume. This is in contrast...