ORIGINAL PAPER

How immunogenetically different are domestic pigs from wildboars: a perspective from single-nucleotide polymorphismsof 19 immunity-related candidate genes

Shanyuan Chen & Rui Gomes & Vânia Costa & Pedro Santos & Rui Charneca &

Ya-ping Zhang & Xue-hong Liu & Shao-qing Wang & Pedro Bento & Jose-Luis Nunes &

József Buzgó & Gyula Varga & István Anton & Attila Zsolnai & Albano Beja-Pereira

Received: 27 March 2013 /Accepted: 17 June 2013# Springer-Verlag Berlin Heidelberg 2013

Abstract The coexistence of wild boars and domestic pigsacross Eurasia makes it feasible to conduct comparative ge-netic or genomic analyses for addressing how geneticallydifferent a domestic species is from its wild ancestor. To testwhether there are differences in patterns of genetic variabilitybetween wild and domestic pigs at immunity-related genesand to detect outlier loci putatively under selection that mayunderlie differences in immune responses, here we analyzed54 single-nucleotide polymorphisms (SNPs) of 19 immunity-related candidate genes on 11 autosomes in three pairs of wildboar and domestic pig populations from China, Iberian

Peninsula, and Hungary. Our results showed no statisticallysignificant differences in allele frequency and heterozygosityacross SNPs between three pairs of wild and domestic popu-lations. This observation was more likely due to the wide-spread and long-lasting gene flow between wild boars anddomestic pigs across Eurasia. In addition, we detected eightcoding SNPs from six genes as outliers being under selectionconsistently by three outlier tests (BayeScan2.1, FDIST2, andArlequin3.5). Among four non-synonymous outlier SNPs,one from TLR4 gene was identified as being subject to posi-tive (diversifying) selection and three each from CD36,

Electronic supplementary material The online version of this article(doi:10.1007/s00251-013-0718-5) contains supplementary material,which is available to authorized users.

S. Chen :R. Gomes :V. Costa :A. Beja-Pereira (*)Centro de Investigação em Biodiversidade e Recursos Genéticos daUniversidade do Porto (CIBIO/UP), Campus Agrário de Vairão,Rua Padre Armando Quintas 7, 4485-661 Vairão, Portugale-mail: albanobp@fc.up.pt

P. Santos :R. Charneca : J.<L. NunesICAAM – Instituto de Ciências Agrárias e AmbientaisMediterrânicas, Universidade de Évora, Herdade da Mitra,7002-554 Évora, Portugal

Y.<p. ZhangState Key Laboratory of Genetic Resources and Evolution,Kunming Institute of Zoology, Chinese Academy of Sciences,Kunming 650223, China

Y.<p. ZhangLaboratory for Conservation and Utilization of Bio-resource,Yunnan University, Kunming 650091, China

X.<h. Liu : S.<q. WangFaculty of Animal Science and Technology, Yunnan AgriculturalUniversity, Kunming 650201, China

P. BentoAssociação Nacional de Criadores de Porco Alentejano (ANCPA),7350 Elvas, Portugal

J. Buzgó :G. VargaSEFAG Forest Management and Wood Industry Share Company,Kaposvár, Hungary

I. Anton :A. ZsolnaiResearch Institute for Animal Breeding and Nutrition,2053 Herceghalom, Hungary

A. ZsolnaiUniversity of Kaposvár, 7400 Kaposvár, Hungary

ImmunogeneticsDOI 10.1007/s00251-013-0718-5

IFNW1, and IL1B genes were suggested as under balancingselection. All of these four non-synonymous variants werepredicted as being benign by PolyPhen-2. Our results weresupported by other independent lines of evidence for positiveselection or balancing selection acting on these four immunegenes (CD36, IFNW1, IL1B, and TLR4). Our study showed anexample applying a candidate gene approach to identify func-tionally important mutations (i.e., outlier loci) in wild anddomestic pigs for subsequent functional experiments.

Keywords Wild boar . Domestic pig . Single nucleotidepolymorphism . Positive selection . Balancing selection

Introduction

How many genetic changes are needed for the transformationfrom a wild into a domestic species? The search for identify-ing such genetic changes that differentiate domestic animalsfrom their wild ancestors and an understanding of the evolu-tionary processes involved have increasingly become one ofthe major research themes in the fields of animal genetics andbreeding. As a result, a growing number of genes and genomicregions have been linked to a variety of phenotypic differ-ences between domestic species and their wild progenitors(e.g., Axelsson et al. 2013; Rubin et al. 2010; vonHoldt et al.2010). Unlike other domestic livestock species whose wildprogenitors are either extinct (domestic cattle and horse) oruncertain (domestic sheep), both domestic pig and its wildcounterpart (i.e., wild boar) are co-distributed across a vastterritory of Eurasia. The coexistence of wild and domestic pigsmakes it feasible to conduct comparative genetic or genomicanalyses for addressing how genetically different are domesticpigs from wild boars.

Generally, compared to their wild counterparts, domesticpigs and other domestic animal species as well have beensubjected to intense selection under domestication and captivebreeding and thus resulted in marked changes in morphologi-cal, physiological, and behavioral traits (Price 2002). For in-stance, domestic pigs have undergone changes in skeletalmuscle growth and cellularity (Rehfeldt et al. 2008), coat color(Andersson 2009), and perhaps disease susceptibility and re-sistance. To dissect the genetic basis underlying differences inphenotypic traits (including disease susceptibility and resis-tance) between wild and domestic pigs or among differentdomestic pig breeds, two general approaches, either top-down (from phenotype to genotype) or bottom-up (from ge-notype to phenotype), have been widely used to discover thecausal genes and genomic regions for monogenic and complextraits (Ross-Ibarra et al. 2007). Although there are severalsuccessful cases using the top-down approach in domestic pigs(see Andersson (2009)), recent studies tend to use the bottom-up approach (also known as genome-wide scan), benefiting

from technological advances in genome sequencing andgenotyping. Recent genome-wide scans in wild and domesticpigs revealed a strong signature of selection in genes mainlyassociated with coat color, morphological changes, productiontraits, olfaction, and immune response (Amaral et al. 2011;Groenen et al. 2012; Rubin et al. 2012; Wilkinson et al. 2013).However, compared to humans, less immunity-related genes inwild and domestic pigs have been detected as targets of posi-tive selection (Groenen et al. 2012) given that they are exposedto different pathogenic environments, for example, virus loads(Reiner et al. 2010).

Indeed previous comparative genomics studies have dem-onstrated that genes involved in immune defenses tend toundergo adaptive evolution at the interspecies level (e.g.,Kosiol et al. 2008). Recent genome-wide scans in humanshave also detected a number of immunity-related genes thatpresent a genomic signature of positive selection (see Akey(2009)). A general interpretation for such patterns of selectionin immunity-related genes is host–pathogen arms race orpathogen-driven selection (Barreiro and Quintana-Murci2010), which can take forms of positive selection (i.e., favoringnew advantageous mutations) and balancing selection (i.e.,favoring genetic variability within a population). Althoughbalancing selection is not a significant force in human genomeas a whole (Andrés et al. 2009), it is a key force for a number ofgenes, including immunity-related genes (Ferrer-Admetllaet al. 2008; Fumagalli et al. 2009). Notably, evidence forinstances of balancing selection is still increasingly mounting(e.g., Leffler et al. 2013). Collectively, exploring genetic var-iation and detecting signatures of selection in immunity-relatedgenes are useful and informative for linking functionally im-portant mutations to disease susceptibility and resistance.

In this study, we applied a candidate gene approach to detectoutlier loci putatively under selection from a panel of innateimmunity-related candidate genes between wild and domesticpig populations using single-nucleotide polymorphism (SNP)data. The populations of choice from China, Iberian Peninsula,and Hungary cover both previously defined Asian and Euro-pean lineages (Giuffra et al. 2000; Megens et al. 2008) andmultiple geographic centers of pig domestication (Larson et al.2005). Interestingly, previous studies have revealed differentpolymorphic patterns in few innate immune genes betweenwild boar and domestic pig populations (e.g., Bergman et al.2010). These pilot results implicate that it has great potential toidentify genes and mutations underlying differences in im-mune responses between wild and domestic pigs by using agenome-wide candidate gene approach.

The main goals of this study were to (1) test whether therewere differences in patterns of genetic variability between threepairs of wild and domestic pig populations at immunity-relatedcandidate genes and (2) detect outlier loci putatively underselection that may underlie differences in immune responsesbetween wild and domestic pigs.

Immunogenetics

Materials and methods

Samples and genomic DNA extraction

A 16-sample panel was initially used in PCR amplificationand sequencing for SNP discovery. The 16-sample panelconsisted of four individuals from China (two wild and twolocal domestic), six individuals from Iberian Peninsula (fourwild and two Alentejano), and six individuals from Hungary(two Mangalica, one Large White, one Landrace, and twoPietrain). A total of 382 individuals (192 domestic pigs and190 wild boars) from China, Iberian Peninsula (Portugal andSpain), and Hungary were used for SNP genotyping. Thepopulation samples included 29 Chinese local domestic pigs(CHD) and 25 Chinese wild boars (CHW) from Yunnanprovince, southwest China; 80 Iberian domestic pigs (IBD,Alentejano from Portugal) and 97 Iberian wild boars (IBW,70 from Portugal and 27 from Spain); 83 HungarianMangalica pigs (HUD, including three Mangalica from Ro-mania) and 68 Hungarian wild boars (HUW). Efforts weremade to avoid sampling related individuals. Genomic DNAwas extracted from ear skin tissues by using JETQUICKTissue DNA Spin Kit (GENOMED GmbH, Lohne, Germa-ny) according to the manufacturer's instructions.

PCR amplification and sequencing

A total of 19 sequence fragments from 13 immunity-relatedcandidate genes (CD40, FASLG, FUT1, IFNG, IFNW1, IL1B,IL6, IL10, IL16, LTA, SLC11A1, TLR2, and TLR4) wereamplified and sequenced. The PCR primers (SupplementaryTable S1) were designed by Primer3 v0.4.0 online web plat-form (http://frodo.wi.mit.edu/) based on NCBI porcine ge-nome assembly Sscrofa9.2 reference sequences. Polymerasechain reaction (PCR) system and working conditions were asdescribed previously (Chen et al. 2011) but with differentannealing temperatures upon primer pairs. PCR products werepurified and sequenced for both strands at the High-Throughput Genomics Unit, Department of Genome Sci-ences, University of Washington (http://www.htseq.org/).The resulting raw sequences were checked and aligned usingsoftware package DNASTAR v7.1 (DNASTAR Inc., Madi-son, WI, USA) to identify SNPs. From those 19 sequencefragments, 70 SNPs and one 9-bp deletion were detected inthe 16-sample panel (Supplementary Table S2).

SNP panel and genotyping

All SNPs identified in this study (Supplementary Table S2),together with 19 SNPs retrieved from six candidate genes(CD36, FOS, GBP2, ITGB1, LTB, and TLR5) in pig SNPdatabases—NCBI or Ensembl (Supplementary Table S3),comprised a set of 90 SNPs, which were used to generate

multiplexed assays. Through the assay design process, 64 outof 90 SNPs were sorted into three iPLEX assays (each with28, 21, and 15 SNPs) by the MassARRAY® Designer soft-ware. These 64 SNPs were composed of 47 coding (includingone 9-bp deletion) and 17 noncoding SNPs from 19 candidategenes on 11 autosomes (Supplementary Table S3).

SNP genotyping was done using SequenomMALDI-TOF-based platform with iPLEXGold technology (Sequenom Inc.,San Diego, CA, USA), following the manufacturer's instruc-tions, at the Genomics Unit of Instituto Gulbenkian de Ciência(http://www.igc.gulbenkian.pt/). The resulting raw genotypesdata were imported into the SNPator web-based platform(Morcillo-Suarez et al. 2008) for quality control and filtering.

Statistical data analyses

Allele frequency and diversity measures (including observedand expected heterozyogsity) were calculated in GenAlEx 6.5(Peakall and Smouse 2012). To explore and visualize thedifferences and relationships between individuals and popu-lations, we carried out principal coordinate analyses (PCoA)in GenAlEx 6.5 using Codom-Genotypic distance—a set ofsquared distances between genotypes (Smouse and Peakall1999). We also calculated population pairwise Fst distancesby Codom-Allelic distance via analysis of molecular variance(AMOVA) option in GenAlEx 6.5, which brings the estimatesof Fst in line with Weir and Cockerham (1984) estimates. Inaddition, to further detect population genetic structure, we ranStructure 2.3.3 (Pritchard et al. 2000) using the default condi-tions of an admixture model and correlated allele frequenciesamong populations (Falush et al. 2003), with a running lengthperiod of 500,000 per run after a burn-in period of 50,000. Therange of the numbers of clusters (K) was tested from one to tenwith three replicates for each K. The Evanno method (Evannoet al. 2005), implemented by STRUCTURE HARVESTER(Earl and vonHoldt 2012), was applied to identify the correctnumber of clusters that best fit the data. The programDISTRUCT 1.1 (Rosenberg 2004) was finally utilized tographically display the population structure.

To detect outlier SNPs potentially under positive(diversifying) selection or balancing selection, we conductedthree Fst-based outlier tests including BayeScan (Foll andGaggiotti 2008), FDIST2 (Beaumont and Nichols 1996),and Arlequin3.5 (Excoffier et al. 2009). The BayeScan ap-proach implements a Bayesianmethod to directly calculate theposterior probability for each locus under the model includingselection; the Posterior Odds (PO)—the ratio of posteriorprobabilities—was used to indicate how more likely the mod-el with selection is compared to the neutral model (Foll andGaggiotti 2008). We ran BayeScan 2.1 using the recommend-ed default parameter settings. The R function “plot_bayescan”was applied to identify outliers using the q-value threshold,leading to a false discovery rate (FDR) of 5 or 1 %.

Immunogenetics

The FDIST2 approach was proposed to obtain the nulldistribution of Fst versus expected heterozygosity under anisland model by coalescent simulations and to identify thoseloci with excessively high or low Fst compared to neutralexpectations as outliers being putatively under selection(Beaumont and Nichols 1996). We performed the FDIST2approach in LOSITAN (Antao et al. 2008) with 100,000simulations and the option of “neutral mean Fst” checked.The confidence interval was set at 0.99, which was suggestedto adjust false-positive rates (Beaumont and Balding 2004).

The Arlequin3.5 approach is essentially an extension of theFDIST2 approach, with inclusion of hierarchical geneticstructure and some differences in other aspects (Excoffieret al. 2009; Excoffier and Lischer 2010). The finite islandmodel in the FDIST2 approach has been recently shown toresult in a large fraction of false-positives under the conditionsthat population samples are hierarchically structured or somepopulation samples have a recent shared history, whereas thehierarchical islandmodel in Arlequin3.5 can reduce the excessof false-positives under such conditions (Excoffier et al.2009). We conducted 50,000 coalescent simulations with 50groups of 100 demes under the hierarchical island models inArlequin3.5.1.3 (Excoffier and Lischer 2010) to generate thejoint distribution of Fst versus heterozygosity. Pre-definedpopulation groupings were set as two groups (Chinese vs.European) or three groups (Chinese vs. Iberian vs. Hungari-an). Loci that fall out of the 99 % confidence intervals of thedistribution were identified as outliers being putatively underselection.

Predicting effects of non-synonymous variants on proteinfunction

We used the software tool Polymorphism Phenotyping v2(PolyPhen-2) (http://genetics.bwh.harvard.edu/pph2/) to pre-dict the functional effects of coding non-synonymous variants.PolyPhen-2 is an automatic probabilistic classifier that uses anumber of sequence- and structure-based predictive featuresby an iterative greedy algorithm (Adzhubei et al. 2010). Weran the PolyPhen-2 web interface (v2.2.2r398) under defaultHumDiv model, which uses 5 or 10 % false-positive ratethresholds for “probably damaging” or “possibly damaging”predictions. For this study, variants were predicted as “be-nign” and “damaging” (including probably and possiblydamaging).

Results and discussion

SNP filtering and data quality

Extensive quality control was performed to assess SNP dataquality based on several criteria including non-concordant

repetitions, negative controls, genotypes validation, andgenotyping call success rate. Initially, throughout 64 SNPsgenotyped, we did not find any SNPs with more than twoalleles (i.e., all genotypes are validated) but found thateight were uninformative (CD36 g.40803A>G, CD40g.10460_10468del, FOS g.2195A>T, FUT1 g.955C>T,FUT1 g.1300G>A, IL1B g.3971C>T, ITGB1 g.40930 T>G,and SLC11A1 g.139C>T). Moreover, there were no discrepantrepetitions across all genotypes examined. Furthermore, whenapplying a cutoff genotyping call success rate of >90 %, wefiltered out two SNPs (CD40 g.10688C>T and IL1Bg.2684G>A) and seven samples, yielding a genotype matrixof 54 SNPs and 375 samples. In addition, five samples ofChinese domestic pigs found to be recent hybrids werealso not considered. Eventually, we obtained a genotypic dataset of 54 SNPs and 370 samples of high quality (Supplemen-tary Table S4) for subsequent population genetic analyses.The final SNP data set contained 39 coding (13 non-synonymous) and 15 noncoding SNPs of 19 candidate genes(CD36, CD40, FASLG, FOS, FUT1, GBP2, IFNG, IFNW1,IL1B, IL6, IL10, IL16, ITGB1, LTA, LTB, SLC11A1, TLR2,TLR4, and TLR5) on 11 different autosomes.

Pairwise comparisons of allele frequency and heterozygositybetween wild and domestic pigs

Overall, Chinese populations (CHW and CHD) showedhigher levels of genetic variability than their counterpartsfrom Iberian Peninsula (IBW and IBD) and Hungary (HUWand HUD) across all diversity measures (Table 1). To testwhether there were differences in patterns of genetic vari-ability between wild and domestic pig populations, weperformed pairwise comparisons of allele frequency andobserved and expected heterzygosity across loci and foundno statistically significant differences (P>0.05, Mann–Whitney U-test) between three pairs of populations fromthree regions (Table 2). When pooling populations from twoEuropean regions together, we also did not find statisticallysignificant differences between wild and domestic pigs(Table 2). It deserves to explore several possibilities forinterpreting such patterns of no differences in allele frequencyand heterozygosity across loci between three pairs of wild anddomestic pig populations being different from those polymor-phic patterns observed in three toll-like receptor genes (Berg-man et al. 2010).

First, this pattern might be attributed to SNP ascertain-ment bias toward common alleles being shared and non-segregated between wild and domestic pig populations. Thisphenomenon was commonly observed in previous studies ongenome-wide SNP variation (e.g., The InternationalHapMap Consortium 2005). To assess this possibility, wecalculated the average minor allele frequency (MAF) acrossall 54 SNPs and obtained an average value of 16.98 %,

Immunogenetics

reflecting SNP ascertainment bias not toward common al-leles. When looking into details at the locus level, we detect-ed 31 SNPs with MAF less than 10 % (accounting for57.41 % of total SNPs), of which 25 SNPs actually possessedMAF below 5 %. These results indicated that the SNP lociwith rare alleles were in fact not underrepresented in thisstudy, and thus estimates of variation within and betweenpopulations would not suffer strong biases. Since rare allelescontribute little to heterozygosity (e.g., Cornuet and Luikart1996), the observed similar level of heterozygosity betweenthree pairs of wild and domestic pig populations was lesslikely due to SNP ascertainment bias.

Second, complex demography could theoretically createsuch observed similar polymorphic patterns between threepairs of wild and domestic pig populations. However, thisexplanation requires complex demographic scenarios suchthat three wild boar populations have undergone contractionsin parallel, whereas three domestic pig populations haveexperienced expansions simultaneously, after their domesti-cation splits (approximately 10,000 years ago). In contrast,previous genetic studies showed that European wild boarsexcept for Italian populations have undergone postglacialexpansions, not contractions (Alves et al. 2010; Scanduraet al. 2008). As a result, this explanation is practically unre-alistic and easily excluded, considering varying spatial andtemporal conditions across China, Iberian Peninsula, andHungary.

Finally, one more likely and parsimonious explanation isextensive hybridization events (i.e., gene flow) between wild

and domestic populations after domestication. This scenariois well supported by our observations. The first observationis lower Fst values between wild and domestic populationsfrom the same geographic region (e.g., 0.071 from China and0.085 from Iberian Peninsula) than those among regions(Table 3), suggesting weak differentiation. The second ob-servation is that evidence for certain levels of gene flowbetween wild and domestic pigs has also been revealed bymitochondrial DNA, microsatellites, and nuclear DNA se-quence makers at different geographic regions, includingChina (Ji et al. 2011), Iberian Peninsula (van Asch et al.2012), and central Europe (Larson et al. 2005; Scandura et al.2008). In addition, a recent study using genome-wide SNPsalso revealed genetic introgression from domestic pigs intowild boars in Northwest Europe (Goedbloed et al. 2013). Thewidespread and long-lasting gene flow between wild anddomestic pigs across Eurasia would homogenize their genepools (Ramírez et al. 2009), leading to a homogenization ofallele frequencies and a similar level of genetic variability.Actually, such historical admixture may have promoted theexchange of advantageous alleles at these immunity-relatedcandidate genes that are nowadays shared between wild anddomestic pig populations.

Population structure of wild and domestic pigs

The PCoA results revealed that the first two axes accountedfor 60.41 and 83.03 % of the total variation at individual(Fig. 1a) and population (Fig. 1b) levels, respectively. The

Table 1 Diversity measures atthe population level

N sample size, %P percentage ofpolymorphic loci, Na number ofdifferent alleles, SE standard er-ror, Ho observed heterozygosity,He expected heterozygosity

Population Number of individuals, N %P Na (SE) Ho (SE) He (SE)

CHW 25 83.33 % 1.833 (0.051) 0.310 (0.041) 0.255 (0.026)

CHD 21 74.07 % 1.741 (0.060) 0.320 (0.042) 0.270 (0.027)

IBW 95 48.15 % 1.481 (0.069) 0.227 (0.045) 0.176 (0.030)

IBD 80 51.85 % 1.519 (0.069) 0.214 (0.044) 0.170 (0.028)

HUW 66 44.44 % 1.444 (0.068) 0.183 (0.043) 0.144 (0.027)

HUD 83 59.26 % 1.593 (0.067) 0.220 (0.042) 0.183 (0.027)

Table 2 Mann–Whitney U-test results of pairwise comparisons between wild boars and domestic pigs

Pair of populations Number of SNPs Allele frequency Ho He

U P U P U P

CHW vs. CHD 48 4,608.0 0.500 1,171.5 0.444 1,167.5 0.455

IBW vs. IBD 31 1,922.5 0.499 496.5 0.411 500.0 0.395

HUW vs. HUD 34 2,312.0 0.501 693.5 0.078 690.0 0.085

(IBW+HUW) vs. (IBD+HUD) 40 3,200.5 0.500 869.5 0.252 852.5 0.306

All P (one-tailed) values are≥0.05; for allele frequency comparisons, n1 and n2 are twice the number of SNPs, whereas forHo andHe comparisons, n1and n2 are equal to the number of SNPs

Ho observed heterozygosity, He expected heterozygosity

Immunogenetics

plot of all individuals on the first two axes showed very lowgenetic identity between Chinese wild and domestic popula-tions and their European counterparts, but with a certainlevel of genetic identity between Chinese (CHD) and Hun-garian (HUD) domestic pigs and also high genetic identitybetween Iberian (IBD) and Hungarian (HUD) domestic pigs(Fig. 1a). Likewise, plot of six populations on the first axisindicated a clear distinction between Chinese and Europeanpopulations, while plot on the second axis revealed a furtherseparation of Hungarian wild boar (HUW) from other threeEuropean populations (Fig. 1b).

Moreover, STRUCTURE analyses obtained essentiallysimilar results as PCoA analyses. The Evanno method(Evanno et al. 2005) identified three clusters (K=3) that bestfit the data (Supplementary Fig. S1). The STRUCTUREresults showed a major split with a limited level of admixturebetween Chinese and European populations, even when Kranged from 2 to 4 (Fig. 2). When K=3, besides separationfrom Chinese counterparts, all European populations eachcontained individuals belonging to two European clusters(Fig. 2), probably reflecting a common descent.

These results, taken together, are consistent with previousgenetic studies that detected two major phylogenetic groupseach for Asian and European wild and domestic pigs (Giuffraet al. 2000; Megens et al. 2008). Albeit a small number ofSNPs used, a certain level of genetic identity between Chi-nese and Hungarian domestic pigs is still discernable. This isbetter interpreted by genetic introgression from Chinese intoEuropean domestic pigs, as clearly demonstrated in recentgenome sequencing and SNP chip genotyping studies(Amaral et al. 2011; Groenen et al. 2012; Wilkinson et al.2013). However, considering a recent genome-wide SNPsstudy which reported that Spanish Iberian pigs are signifi-cantly differentiated from Romanian Mangalica pigs(Manunza et al. 2013), high genetic identity between Iberian

Table 3 Population pairwise Fst distances by GenAlEx Codom-Allelicdistance via AMOVA

Population CHW CHD IBW IBD HUW HUD

CHW 0.000

CHD 0.071 0.000

IBW 0.260 0.305 0.000

IBD 0.300 0.313 0.085 0.000

HUW 0.338 0.364 0.148 0.161 0.000

HUD 0.246 0.270 0.086 0.040 0.135 0.000

All P values are≤0.001 based on 9,999 permutations

Axi

s 2

(26.

01%

)

Axis 1 (34.40%)

CHW

CHD

IBW

IBD

HUW

HUD

CHW

CHDIBW

IBD

HUW

HUD

Axi

s 2

(14.

13%

)

Axis 1 (68.90% )

a

b

Fig. 1 Principal coordinatesanalyses (PCoA) of GenAlExCodom-Genotypic distances atindividual and population levels.a Plot of axis 1 against axis 2 atindividual level. b Plot of axis 1against axis 2 at population level

Immunogenetics

(Alentejano from Portugal) and Hungarian Mangalica pigs inthis study tends to be attributed to a common descent, not toadmixture, despite the fact that a panel of 54 SNPs has alimited resolution.

Detection of outlier SNPs being putatively under selection

When considering all populations cumulatively, BayeScan2.1revealed eight outlier SNPs (four non-synonymous) with pos-terior probabilities (PP) ranging from 0.978 to 1.0 at a FDR of5 %, including two for positive (diversifying) selection and sixfor balancing or purifying selection, as indicated by alphavalues (Table 4). When adjusting to a FDR of 1 %, three non-

synonymous SNPs (IFNW1 g.391G>A, IL1B g.4108C>A,and TLR4 g.7485C>A) remained as outliers (those with q-value<0.01). Noticeably, BayeScan2.1 gave decisive evidence(PP=1.0) for positive (diversifying) selection acting on TLR4g.7485C>A, which causes an amino acid change Lys343Gln.In addition, FDIST2 (LOSITAN) detected ten outlier SNPs(Fig. 3), including those eight by BayeScan2.1 and two newones (IFNG g.1703 T>C and ITGB1 g.30721G>A). It is worthnoting that IFNG g.1703 T>C, suggested by FIDIST2 as underpositive (diversifying) selection, may indeed represent a false-positive given its noncoding status and low Fst value (< 0.05).

We were aware that both BayeScan2.1 and FDIST2methods do not take hierarchical population structures into

CH

W

CH

D

IBW

IBD

HU

W

HU

D

K = 4

K = 3

K = 2

Fig. 2 DISTRUCT plots of STRUCTURE results at K=2, 3, and 4, respectively

Table 4 Outlier SNPs putativelyunder selection as detected byBayeScan2.1

a All are coding SNPs with non-synonymous ones shown in italicb Posterior probability for themodel including selectionc Coefficient indicating the strengthand direction of selection (i.e., pos-itive values—diversifying selectionand negative values—balancing orpurifying selection)

Gene SNPa Probb Log10(PO) q-value Alphac Fst

CD36 g.40748G>A 0.983 1.767 0.012 −2.110 0.071

CD36 g.40827A>G 0.981 1.708 0.013 −2.108 0.071

FOS g.2086A>G 0.978 1.656 0.015 1.370 0.598

GBP2 g.13067A>G 0.980 1.681 0.014 −2.047 0.074

IFNW1 g.391G>A 0.986 1.860 0.007 −2.092 0.072

IL1B g.2744C>T 0.985 1.817 0.011 −2.112 0.071

IL1B g.4108C>A 0.986 1.848 0.009 −2.116 0.070

TLR4 g.7485C>A 1.000 ∞ 0.000 1.864 0.697

Immunogenetics

account, and our studied populations with geographic struc-turing as revealed by PCoA and STRUCTURE analysesviolated their assumptions. To reduce a confounding effectof geographic structuring among populations, we thusconducted Arlequin3.5 analyses under hierarchical islandmodels. When setting at two groups (Chinese vs. European),Arlequin3.5 found exactly the same eight outliers as byBayeScan2.1 (Fig. 4), while when setting at three groups(Chinese vs. Iberian vs. Hungarian), Arlequin3.5 identified11 outliers, including those aforementioned eight and threenew ones (GBP2 g.12934C>A, GBP2 g.12936C>A, andIFNW1 g.435C>T) (Fig. S2). Moreover, as recent studiesdemonstrated, all of the three outlier tests applied here maysuffer type I (false-positive) and type II (false-negative) errors;it is important to interpret outliers cautiously (e.g., Narum andHess 2011). We thus only considered those eight outlier SNPs(Table 4; Fig. 4) consistently revealed by all three methods ascandidate loci targeted by selection.

We also acknowledged that Fst-based outlier tests typicallyneed hundreds to thousands of SNPs to build a robust nullneutral distribution of Fst (e.g., Beaumont and Balding 2004;Excoffier et al. 2009). Our study using a panel of 54 SNPswould expectedly have limited statistical power to preciselytell outliers apart from neutral loci. However, it should beemphasized that, though small in number, our SNPs are alllocated in immunity-related genes belonging to top categoriesof genes that have been often found under positive(diversifying) or balancing selection in recent genome-widestudies (e.g., Akey 2009; Kosiol et al. 2008). Thus, it is stilllikely to detect few outliers putatively under selection fromthis small particular data set, as demonstrated through our

analyses that revealed eight outliers constantly across threedifferent methods.

Furthermore, it is known that purifying selection againstslightly deleterious non-synonymous variants would also in-crease or decrease levels of population differentiation, con-founding to effects of positive selection or balancing selection(e.g., Barreiro et al. 2008; Bustamante et al. 2005). However,knowing the functional effects of non-synonymous variantswould enable one to distinguish between different types ofselection. By using PolyPhen-2, we found that the four non-synonymous outlier SNPs were predicted as “benign”(Table 5), suggesting that purifying selection seems less likelyto account for their exhibited levels of population differentia-tion (Table 4). Thus, the high Fst value of one outlier (TLR4g.7485C>A) and the low Fst values of three outliers (CD36g.40827A>G, IFNW1 g.391G>A, and IL1B g.4108C>A)were more likely to be caused by positive (diversifying)selection and balancing selection, respectively. Regardingthose four synonymous outlier SNPs (Table 4), it was under-standable through hitchhiking effect on linkage disequilibriumbetween linked neutral loci (Stephan et al. 2006). This wasobviously true for two synonymous outliers (CD36g.40748G>A and IL1B g.2744C>T), each having one nearbynon-synonymous outlier SNP detected as under selection asmentioned earlier.

Always bear in mind that, for outlier loci inferred as underselection, we need to understand their selective agents ormotivation for selection beyond statistical tests. As mentionedearly on, a general interpretation for positive (diversifying)selection or balancing selection acting on immunity-relatedgenes is pathogen-driven selection (Barreiro and Quintana-

Fig. 3 The outliers suggested by FDIST2 method implemented in LOSITAN

Immunogenetics

Murci 2010). This explanation may also be applied to thosefour non-synonymous outlier SNPs of selection signaturesidentified in wild and domestic pigs here. To gain moresupport for this explanation, we examined the evolutionary

patterns of these four immune genes (CD36, IFNW1, IL1B,and TLR4) from other independent studies. Previous geneticstudies in human populations showed that a CD36 nonsensevariant, associated with differential susceptibility to cerebralmalaria, had undergone recent positive selection (see Sabetiet al. (2006)). Evidence (albeit weak) for positive selectionwas detected in cytokine IFNW1 gene in human populations(Amos and Bryant 2011). Interestingly, strong evidence forbalancing selection acting on cytokine IL1B gene (same as inour study) has been found in a natural population of field voles(Turner et al. 2012). Most importantly, genetic variation of theIL1B gene had already been demonstrated to be associatedwith immune variation and resistance to multiple pathogens infield voles (Turner et al. 2011). In addition, several studiesdemonstrated that TLR4 gene—one of most well-known pat-tern recognition receptors, had been subject to positive selec-tion, for example, in bovine species (White et al. 2003) and inprimates (Wlasiuk and Nachman 2010). These independentlines of evidence gave clear support for our interpretation thatthose four non-synonymous outlier SNPs may have beenunder pathogen-mediated selection.

It was worthy of notice that we detected four out of sixgenes harboring outlier SNPs (or four out of 19 genes stud-ied) as being putatively under balancing selection. This

Table 5 The PolyPhen-2 predictions for 13 non-synonymous variants

Gene SNPa Substitution Scoreb Prediction

CD36 g.40827A>G Lys86Glu 0.000 Benign

FUT1 g.1074C>A Ala155Asp 0.620 Damaging

GBP2 g.12934C>A Ala219Glu 0.000 Benign

GBP2 g.12936C>A Pro220Thr 0.745 Damaging

IFNW1 g.391G>A Ala131Thr 0.000 Benign

IL1B g.4108C>A Thr147Lys 0.079 Benign

IL6 g.323G>T Arg31Leu 0.003 Benign

SLC11A1 g.23G>A Val175Ile 0.000 Benign

TLR4 g.7420G>A Arg321His 0.000 Benign

TLR4 g.7485C>A Gln343Lys 0.000 Benign

TLR5 c.834 T>G His278Gln 0.000 Benign

TLR5 c.902C>T Ser301Phe 0.999 Damaging

TLR5 c.959 T>A Phe320Tyr 0.000 Benign

a The SNPs detected as outliers under selection shown in italicb The score is posterior probability of a mutation being damaging

Heterozygosity/(1-Fst)

Fst

TLR4 g.7485C>A

FOS g.2086A>G

IL1B g.2744C>TIL1B g.4108C>A

IFNW1 g.391G>ACD36 g.40748G>A

GBP2g.13067A>GCD36 g.40827A>G

Fig. 4 The outliers identified byArlequin3.5 hierarchical islandmodel at two groups (Chinese vs.European)

Immunogenetics

proportion appeared too high at first sight, when comparingto 60 out of 13,400 genes as targets of balancing selectionacross human genome (Andrés et al. 2009). However, whencomparing to those studies that concentrated only on im-mune genes, we found a similar proportion of genes identi-fied as under balancing selection. For example, Ferrer-Admetlla et al. (2008) found signatures of balancing selec-tion in six out of nine immune genes analyzed; Fumagalliet al. (2009) demonstrated that three out of 26 blood groupantigen genes had been subject to balancing selection. Al-though it is difficult to detect balancing selection using justoutlier tests (Hofer et al. 2012), these three non-synonymousoutlier SNPs (CD36 g.40827A>G, IFNW1 g.391G>A, andIL1B g.4108C>A) should represent true targets of balancingselection in wild and domestic pigs. This is particularly truefor the outlier IL1B g.4108C>A because there was indepen-dent evidence for balancing selection acting on the IL1Bgene in field voles (Turner et al. 2012). Nonetheless, futurefine-scale population genetic analyses on these three candi-date genes (CD36, IFNW1, and IL1B) by re-sequencing theirfull gene sequences and by integrating appropriate data anal-ysis methods such as haplotype-based and linkagedisequilibrium-based methods (see Sabeti et al. (2006)) arehighly needed. In addition, functional experiments are re-quired to determine the mechanisms of balancing selection(e.g., heterozygote advantage or frequency-dependent selec-tion) in these three genes in wild and domestic pig popula-tions that are exposed to different pathogenic environments.

Conclusions

Although there were no significantly different patterns ofallele frequency and heterozygosity between three pairs ofwild boar and domestic pig populations in this small panel ofSNPs, encouragingly, we could still identify four non-synonymous outlier SNPs from four immune genes (CD36,IFNW1, IL1B, and TLR4), including one under positive(diversifying) selection and three under balancing selection.Our results, together with other independent studies on thosefour genes, are in agreement with a general view thatpathogen-mediated selection has shaped patterns of geneticvariation in immunity-related genes across species (Barreiroand Quintana-Murci 2010). Our study showed an exampleapplying a candidate gene approach to identify function-ally important mutations (i.e., outlier loci) in wild anddomestic pigs for subsequent functional experiments. Nev-ertheless, advances of next-generation sequencing technol-ogies have produced whole-genome sequencing data with-out ascertainment bias that identify more functionally im-portant mutations underlying phenotypic differences inwild boar and domestic pigs (Groenen et al. 2012; Rubinet al. 2012).

Acknowledgments We thank two anonymous reviewers for theirconstructive comments. This work was financially supported byFundação para a Ciência e Tecnologia (FCT) projects PTDC/CVT/68907/2006 (ABP) and PTDC/CVT/099782/2008 (SC). SC is support-ed by FCT individual grant SFRH/BPD/46082/2008.

Conflicting interest The authors declare that they have no competinginterests.

References

Adzhubei IA, Schmidt S, Peshkin L, Ramensky VE, Gerasimova A, BorkP, Kondrashov AS, Sunyaev SR (2010) A method and server forpredicting damaging missense mutations. Nat Methods 7:248–9

Akey JM (2009) Constructing genomic maps of positive selection inhumans: where do we go from here? Genome Res 19:711–722

Alves PC, Pinheiro I, Godinho R, Vicente J, Gortazar C, Scandura M(2010) Genetic diversity of wild boar populations and domesticpig breeds (Sus scrofa) in south-western Europe. Biol J Linn Soc101:797–822

Amaral AJ, Ferretti L, Megens H-J, Crooijmans RPMA, Nie H, Ramos-Onsins SE, Perez-Enciso M, Schook LB, Groenen MAM (2011)Genome-wide footprints of pig domestication and selection re-vealed through massive parallel sequencing of pooled DNA. PLoSOne 6:e14782

Amos W, Bryant C (2011) Using human demographic history to infernatural selection reveals contrasting patterns on different familiesof immune genes. Proc R Soc B Biol Sci 278:1587–1594

Andersson L (2009) Studying phenotypic evolution in domestic ani-mals: a walk in the footsteps of Charles Darwin. Cold Spring HarbSymp Quant Biol 74:319–325

Andrés AM, Hubisz MJ, Indap A, Torgerson DG, Degenhardt JD,Boyko AR, Gutenkunst RN, White TJ, Green ED, BustamanteCD, Clark AG, Nielsen R (2009) Targets of balancing selection inthe human genome. Mol Biol Evol 26:2755–2764

Antao T, Lopes A, Lopes R, Beja-Pereira A, Luikart G (2008)LOSITAN: a workbench to detect molecular adaptation based ona Fst-outlier method. BMC Bioinforma 9:323

Axelsson E, Ratnakumar A, Arendt M-L, Maqbool K, Webster MT,Perloski M, Liberg O, Arnemo JM, Hedhammar A, Lindblad-TohK (2013) The genomic signature of dog domestication revealsadaptation to a starch-rich diet. Nature 495:360–364

Barreiro LB, Laval G, Quach H, Patin E, Quintana-Murci L (2008)Natural selection has driven population differentiation in modernhumans. Nat Genet 40:340–5

Barreiro LB, Quintana-Murci L (2010) From evolutionary genetics tohuman immunology: how selection shapes host defence genes.Nat Rev Genet 11:17–30

Beaumont MA, Balding DJ (2004) Identifying adaptive genetic diver-gence among populations from genome scans. Mol Ecol 13:969–980

Beaumont MA, Nichols RA (1996) Evaluating loci for use in thegenetic analysis of population structure. Proc R Soc London B263:1619–1626

Bergman I-M, Rosengren J, Edman K, Edfors I (2010) European wildboars and domestic pigs display different polymorphic patterns inthe Toll-like receptor (TLR) 1, TLR2, and TLR6 genes. Immuno-genetics 62:49–58

Bustamante CD, Fledel-Alon A, Williamson S, Nielsen R, Hubisz MT,Glanowski S, Tanenbaum DM, White TJ, Sninsky JJ, HernandezRD, Civello D, Adams MD, Cargill M, Clark AG (2005) Naturalselection on protein-coding genes in the human genome. Nature437:1153–7

Immunogenetics

Chen S, Costa V, Beja-Pereira A (2011) Evolutionary patterns of twomajor reproduction candidate genes (Zp2 and Zp3) reveal nocontribution to reproductive isolation between bovine species.BMC Evol Biol 11:24

Cornuet JM, Luikart G (1996) Description and power analysis of twotests for detecting recent population bottlenecks from allele fre-quency data. Genetics 144:2001–2014

Earl D, vonHoldt B (2012) STRUCTURE HARVESTER: a website andprogram for visualizing STRUCTURE output and implementing theEvanno method. Conserv Genet Resour 4:359–361

Evanno G, Regnaut S, Goudet J (2005) Detecting the number of clustersof individuals using the software structure: a simulation study. MolEcol 14:2611–2620

Excoffier L, Hofer T, Foll M (2009) Detecting loci under selection in ahierarchically structured population. Heredity 103:285–298

Excoffier L, Lischer HEL (2010) Arlequin suite ver 3.5: a new series ofprograms to perform population genetics analyses under Linuxand Windows. Mol Ecol Resour 10:564–567

Falush D, Stephens M, Pritchard JK (2003) Inference of populationstructure using multilocus genotype data: linked loci and correlat-ed allele frequencies. Genetics 164:1567–1587

Ferrer-Admetlla A, Bosch E, Sikora M, Marquès-Bonet T, Ramírez-Soriano A, Muntasell A, Navarro A, Lazarus R, Calafell F,Bertranpetit J, Casals F (2008) Balancing selection is the mainforce shaping the evolution of innate immunity genes. J Immunol181:1315–1322

Foll M, Gaggiotti O (2008) A Genome-Scan Method to Identify Select-ed Loci Appropriate for Both Dominant and Codominant Markers:A Bayesian Perspective. Genetics 180:977–993

Fumagalli M, Cagliani R, Pozzoli U, Riva S, Comi GP, Menozzi G,Bresolin N, Sironi M (2009) Widespread balancing selection andpathogen-driven selection at blood group antigen genes. GenomeRes 19:199–212

Giuffra E, Kijas JMH, Amarger V, Carlborg Ö, Jeon J-T, Andersson L(2000) The Origin of the domestic pig: independent domesticationand subsequent introgression. Genetics 154:1785–1791

Goedbloed DJ, Megens HJ, Van Hooft P, Herrero-Medrano JM, LutzW,Alexandri P, Crooijmans RPMA, Groenen M, Van Wieren SE,Ydenberg RC, Prins HHT (2013) Genome-wide single nucleotidepolymorphism analysis reveals recent genetic introgression fromdomestic pigs into Northwest European wild boar populations.Mol Ecol 22:856–866

Groenen MAM, Archibald AL, Uenishi H, Tuggle CK, Takeuchi Y,Rothschild MF, Rogel-Gaillard C, Park C, Milan D, Megens H-J,Li S, Larkin DM, Kim H, Frantz LAF, Caccamo M, Ahn H, AkenBL, Anselmo A, Anthon C, Auvil L, Badaoui B, Beattie CW,Bendixen C, Berman D, Blecha F, Blomberg J, Bolund L, BosseM, Botti S, Bujie Z, Bystrom M, Capitanu B, Carvalho-Silva D,Chardon P, Chen C, Cheng R, Choi S-H, ChowW, Clark RC, CleeC, Crooijmans RPMA, Dawson HD, Dehais P, De Sapio F, DibbitsB, Drou N, Du Z-Q, Eversole K, Fadista J, Fairley S, Faraut T,Faulkner GJ, Fowler KE, Fredholm M, Fritz E, Gilbert JGR,Giuffra E, Gorodkin J, Griffin DK, Harrow JL, Hayward A, HoweK, Hu Z-L, Humphray SJ, Hunt T, Hornshoj H, Jeon J-T, Jern P,Jones M, Jurka J, Kanamori H, Kapetanovic R, Kim J, Kim J-H,Kim K-W, Kim T-H, Larson G, Lee K, Lee K-T, Leggett R, LewinHA, Li Y, LiuW, Loveland JE, Lu Y, Lunney JK, Ma J, Madsen O,Mann K, Matthews L, McLaren S, Morozumi T, Murtaugh MP,Narayan J, Truong Nguyen D, Ni P, Oh S-J, Onteru S, Panitz F,Park E-W et al (2012) Analyses of pig genomes provide insightinto porcine demography and evolution. Nature 491:393–398

Hofer T, Foll M, Excoffier L (2012) Evolutionary forces shapinggenomic islands of population differentiation in humans. BMCGenomics 13:107

Ji Y-Q, Wu D-D, Wu G-S, Wang G-D, Zhang Y-P (2011) Multi-locusanalysis reveals a different pattern of genetic diversity for

mitochondrial and nuclear DNA between wild and domestic pigsin East Asia. PLoS One 6:e26416

Kosiol C, Vinař T, da Fonseca RR, Hubisz MJ, Bustamante CD, NielsenR, Siepel A (2008) Patterns of positive selection in six mammaliangenomes. PLoS Genet 4:e1000144

Larson G, Dobney K, Albarella U, Fang M, Matisoo-Smith E, Robins J,Lowden S, Finlayson H, Brand T, Willerslev E, Rowley-Conwy P,Andersson L, Cooper A (2005) Worldwide phylogeography ofwild boar reveals multiple centers of pig domestication. Science307:1618–1621

Leffler EM, Gao Z, Pfeifer S, Ségurel L, Auton A, Venn O, Bowden R,Bontrop R, Wall JD, Sella G, Donnelly P, McVean G, PrzeworskiM (2013) Multiple instances of ancient balancing selection sharedbetween humans and chimpanzees. Science 339:1578–1582

Manunza A, Zidi A, Yeghoyan S, Balteanu VA, Carsai TC, ScherbakovO, Ramírez O, Eghbalsaied S, Castelló A, Mercadé A, Amills M(2013) A high throughput genotyping approach reveals distinctiveautosomal genetic signatures for European and near eastern wildboar. PLoS One 8:e55891

MegensH-J, Crooijmans R, CristobalM, Hui X, Li N, GroenenM (2008)Biodiversity of pig breeds from China and Europe estimated frompooled DNA samples: differences in microsatellite variation be-tween two areas of domestication. Genet Sel Evol 40:103–128

Morcillo-Suarez C, Alegre J, Sangros R, Gazave E, de Cid R, Milne R,Amigo J, Ferrer-Admetlla A, Moreno-Estrada A, Gardner M,Casals F, Pérez-Lezaun A, Comas D, Bosch E, Calafell F,Bertranpetit J, Navarro A (2008) SNP analysis to results(SNPator): a web-based environment oriented to statistical geno-mics analyses upon SNP data. Bioinformatics 24:1643–1644

Narum SR, Hess JE (2011) Comparison of FST outlier tests for SNPloci under selection. Mol Ecol Resour 11:184–194

Peakall R, Smouse PE (2012) GenAlEx 6.5: genetic analysis in Excel.Population genetic software for teaching and research—an update.Bioinformatics 28:2537–9

Price EO (2002) Animal domestication and behavior. CABI, WallingfordPritchard JK, Stephens M, Donnelly P (2000) Inference of population

structure using multilocus genotype data. Genetics 155:945–959Ramírez O, Ojeda A, Tomàs A, Gallardo D, Huang LS, Folch JM, Clop

A, Sánchez A, Badaoui B, Hanotte O, Galman-Omitogun O,Makuza SM, Soto H, Cadillo J, Kelly L, Cho IC, Yeghoyan S,Pérez-Enciso M, Amills M (2009) Integrating Y-chromosome,mitochondrial, and autosomal data to analyze the origin of pigbreeds. Mol Biol Evol 26:2061–2072

Rehfeldt C, Henning M, Fiedler I (2008) Consequences of pig domes-tication for skeletal muscle growth and cellularity. Livest Sci116:30–41

Reiner G, Bronnert B, Hohloch C, Fresen C, Haack I, Willems H,Reinacher M (2010) Qualitative and quantitative distribution ofPCV2 in wild boars and domestic pigs in Germany. Vet Microbiol145:1–8

Rosenberg NA (2004) Distruct: a program for the graphical display ofpopulation structure. Mol Ecol Notes 4:137–138

Ross-Ibarra J, Morrell PL, Gaut BS (2007) Plant domestication, aunique opportunity to identify the genetic basis of adaptation. PNatl Acad Sci U S A 104:8641–8648

Rubin C-J, Megens H-J, Barrio AM, Maqbool K, Sayyab S,Schwochow D, Wang C, Carlborg Ö, Jern P, Jørgensen CB,Archibald AL, Fredholm M, Groenen MAM, Andersson L(2012) Strong signatures of selection in the domestic pig genome.Proc Natl Acad Sci 109:19529–19536

Rubin C-J, Zody MC, Eriksson J, Meadows JRS, Sherwood E, WebsterMT, Jiang L, Ingman M, Sharpe T, Ka S, Hallbook F, Besnier F,Carlborg O, Bed/'hom B, Tixier-Boichard M, Jensen P, Siegel P,Lindblad-Toh K, Andersson L (2010) Whole-genome resequencingreveals loci under selection during chicken domestication. Nature464:587–591

Immunogenetics

Sabeti PC, Schaffner SF, Fry B, Lohmueller J, Varilly P,Shamovsky O, Palma A, Mikkelsen TS, Altshuler D, LanderES (2006) Positive natural selection in the human lineage.Science 312:1614–1620

Scandura M, Iacolina L, Crestanello B, Pecchioli E, Di Benedetto MF,Russo V, Davoli R, Apollonio M, Bertorelle G (2008) Ancient vs.recent processes as factors shaping the genetic variation of theEuropean wild boar: are the effects of the last glaciation stilldetectable? Mol Ecol 17:1745–1762

Smouse PE, Peakall R (1999) Spatial autocorrelation analysis of indi-vidual multiallele and multilocus genetic structure. Hered (Edinb)82(Pt 5):561–73

Stephan W, Song YS, Langley CH (2006) The hitchhiking effect onlinkage disequilibrium between linked neutral loci. Genetics172:2647–2663

The International HapMap Consortium (2005) A haplotype map of thehuman genome. Nature 437:1299–1320

Turner AK, Begon M, Jackson JA, Bradley JE, Paterson S (2011)Genetic diversity in cytokines associated with immune variationand resistance to multiple pathogens in a natural rodent population.PLoS Genet 7:e1002343

Turner AK, Begon M, Jackson JA, Paterson S (2012) Evidence forselection at cytokine loci in a natural population of field voles(Microtus agrestis). Mol Ecol 21:1632–1646

van Asch B, Pereira F, Santos LS, Carneiro J, Santos N, AmorimA (2012)Mitochondrial lineages reveal intense gene flow between Iberianwild boars and South Iberian pig breeds. Anim Genet 43:35–41

vonHoldt BM, Pollinger JP, Lohmueller KE, Han E, Parker HG, QuignonP, Degenhardt JD, Boyko AR, Earl DA, Auton A, Reynolds A, BrycK, Brisbin A, Knowles JC, Mosher DS, Spady TC, Elkahloun A,Geffen E, Pilot M, Jedrzejewski W, Greco C, Randi E, Bannasch D,Wilton A, Shearman J, Musiani M, Cargill M, Jones PG, Qian Z,Huang W, Ding Z-L, Y-p Z, Bustamante CD, Ostrander EA,Novembre J, Wayne RK (2010) Genome-wide SNP and haplotypeanalyses reveal a rich history underlying dog domestication. Nature464:898–902

Weir BS, Cockerham CC (1984) Estimating F-statistics for the analysisof population structure. Evolution 38:1358–1370

White SN, Taylor KH, Abbey CA, Gill CA, Womack JE (2003) Hap-lotype variation in bovine Toll-like receptor 4 and computationalprediction of a positively selected ligand-binding domain. P NatlAcad Sci U S A 100:10364–10369

Wilkinson S, Lu ZH, Megens H-J, Archibald AL, Haley C, Jackson IJ,Groenen MAM, Crooijmans RPMA, Ogden R, Wiener P (2013)Signatures of diversifying selection in European pig breeds. PLoSGenet 9:e1003453

Wlasiuk G, Nachman MW (2010) Adaptation and constraint at toll-likereceptors in primates. Mol Biol Evol 27:2172–2186

Immunogenetics