ebook img

Identification of protein-damaging mutations in 10 swine taste receptors and 191 appetite-reward ... PDF

17 Pages·2016·0.85 MB·English
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Identification of protein-damaging mutations in 10 swine taste receptors and 191 appetite-reward ...

Clopetal.BMCGenomics (2016) 17:685 DOI10.1186/s12864-016-2972-z RESEARCH ARTICLE Open Access Identification of protein-damaging mutations in 10 swine taste receptors and 191 appetite-reward genes Alex Clop1*, Abdoallah Sharaf1,2, Anna Castelló1, Sebastián Ramos-Onsins1, Susanna Cirera3, Anna Mercadé1, Sophia Derdak4,5, Sergi Beltran4,5, Abe Huisman6, Merete Fredholm3, Pieter van As7 and Armand Sánchez1,8* Abstract Background: Taste receptors (TASRs) are essential for thebody’srecognitionof chemical compounds. Inthe tongue, TASRs sense the sweet and umami and the toxin-related bitter taste thus promoting a particular eating behaviour. Moreover, theirrelevance inother organs is now becoming evident. Inthe intestine, theyregulate nutrient absorption and gut motility. Upon ligand binding, TASRs activate theappetite-reward circuitry to signal thenervous system and keep body homeostasis. With the aim to identify genetic variation intheswine TASRs and in thegenes from theappetiteand thereward pathways, we have sequenced theexons of 201TASRsand appetite-reward genes from 304 pigs belonging to ten breeds, wild boars and to two phenotypically extreme groups from a F resource with data on growth and fat deposition. 2 Results: We identified 2,766codingvariants 395 ofwhich were predictedto have a strong impact onprotein sequence and function. 334 variants were present inonly one breed and atpredicted alternativeallele frequency (pAAF)≥0.1. The Asian pigs and thewild boars showed thelargest proportion of breed specific variants. We also compared thepAAF ofthe two F groups and found thatvariants inTAS2R39 and CD36 display significant 2 differences suggesting that these genes could influence growthand fat deposition.We developed a 128-variant genotyping assay and confirmed 57 of thesevariants. Conclusions: We have identified thousandsof variants affecting TASRs as well as genes involved inthe appetite and the reward mechanisms. Some ofthesegenes have been already associated to taste preferences, appetite or behaviour inhumans and mouse. We have also detected indications of a potential relationship of some of these genes withgrowth and fat deposition, which could have been caused by changes intaste preferences, appetite or reward and ultimately impact on food intake. A genotypingarray with57 variantsin 31 of thesegenes is now available for genotyping and start elucidating the impact of genetic variation in thesegenes on pig biology and breeding. Keywords: Taste receptors, Appetite-rewardpathways,Coding genetic variation, Genotyping array, Future association studies *Correspondence:[email protected];[email protected] 1CentreforResearchinAgriculturalGenomics(CRAG)CSIC-IRTA-UAB-UB, CampusUAB,08193CerdanyoladelValles,Catalonia,Spain Fulllistofauthorinformationisavailableattheendofthearticle ©2016TheAuthor(s).OpenAccessThisarticleisdistributedunderthetermsoftheCreativeCommonsAttribution4.0 InternationalLicense(http://creativecommons.org/licenses/by/4.0/),whichpermitsunrestricteduse,distribution,and reproductioninanymedium,providedyougiveappropriatecredittotheoriginalauthor(s)andthesource,providealinkto theCreativeCommonslicense,andindicateifchangesweremade.TheCreativeCommonsPublicDomainDedicationwaiver (http://creativecommons.org/publicdomain/zero/1.0/)appliestothedatamadeavailableinthisarticle,unlessotherwisestated. Clopetal.BMCGenomics (2016) 17:685 Page2of17 Background during fasting also promotes glutamate release. This There are five canonical tastes that are sensed in the neurotransmitter, via a large catalogue of receptors, will taste buds of the tongue, which include salty, sour, also excite NPY, AgRP and inhibit POMC neurons and sweet, umami and bitter. While salty and sour are de- boost appetite. The catabolic product of glutamate, tected by ion channels, the other three are sensed by a gamma aminobutyric acid (GABA) also boosts appetite group of G-protein coupled receptors called taste recep- butusingdifferentneuronalmechanisms[10].Moreover, tors (TASRs). Sweet and umami are appetizing tastes glutamate is also able to excite dopamine-secreting neu- that characterize energy-rich food sources, namely sugar ronsinappetite-relevantareasofthebraintherebyindir- and amino acid molecules, respectively, and are sensed ectly promoting the feeling of hunger [9]. In contrast, by the TAS1R sub-type TAS1R1, TAS1R2 and TAS1R3 GABA can inhibit the same neurons and indirectly pro- [1].Ontheotherhand,the–unpleasant-bittertastein- mote satiety [9]. Serotonin, a neurotransmitter that is dicating the presence of toxic molecules, is sensed by mostly present in the gastrointestinal tract but also at TAS2Rs, also known as bitter taste receptors [1], which much lower levels in the central nervous system, modu- include a variable list of highly polymorphic genes with lates gastrointestinal motility, mood and appetite [11]. many species-specific orthologs. The annotation of the Serotonin inhibits appetite by stimulating its receptors pig genome contains ten TAS2Rs according to the HTR2C and HTR1B, which in turn, activate the well- Ensembl database (www.ensembl.org). In the recent known appetite-inhibitors POMC and MC4R and inhibit years, it has become obvious that TASRs are expressed the appetite-promoter genes NPY and AGRP. Epineph- in many other tissues and have additional chemo- rine/norepinephrine are two additional neurotransmit- sensing functions. For example, they are present in the ters that seem to be key in food intake and in keeping respiratory system where they regulate innate immunity energy balance and have been shown to respond to star- and infection [2], and in sperm they have been linked to vation and low glucose levels in blood by activating the motility and acrosomal reaction [3]. In the gastro- secretion ofghrelin[12]. intestinal tract, TASRs detect the molecules that are on In humans, the recent advent of whole genome [13] transit and stimulate the appetite and reward (AR) cir- and exome [14] sequencing has shown that mutations cuitries to promote the appropriate feeding behaviour, severely impacting on protein sequence are more abun- thus keeping energy balance and body homeostasis [4, dant than previously thought, although due to purifying 5]. The AR mechanisms are highly interconnected and selection, they tend to have very low allele frequencies. involve complex networks containing nutrients, neuro- Thus, protein-damaging polymorphisms in TASRs and peptides, neurotransmitters, hormones and their related AR genes are likely to have an important impact on a receptors and enzymes. These pathways engage the broad range of traits including feed intake, immune gastrointestinal tract, pancreas, liver, muscle, adipose tis- function, behaviour or fertility both in livestock and sue and brain. Appetite-related genes such as leptin humans. Consequently, understanding how these vari- (LEP), leptin receptor (LEPR), cholecystokinin (CCK), ants affect phenotypes in farm animals may both help Ghrelin (GHRL), Agouti-related protein (AgRP), neuro- improving the sustainability of the animal breeding sec- peptide Y (NPY), proopiomelanocortin (POMC) and tor as well as benefit bio-medical research. Scientific melanocortin 4 receptor (MC4R) encode for products interest in this gene family in the pig is now emerging that inhibit or excite the dopamine, epinephrine, nor- and it has recently been shown that TASRs are epinephrine, serotonin, and glutamate receptor pathways expressed in multiple porcine systems (immune, gastro [6, 7] and modulate food intake and energy balance. In a intestinal, spermatogenic, etc. [15, 16]). The recent pub- nutshell, ghrelin and LEP are two hormones with oppos- lication of the swine genome sequence and annotation ite excitatory and inhibitory effects on the same neurons [17] and the development of genome capture assays secreting appetite inducing NPY and AgRP or the feed- open unprecedented possibilities to identify deleterious ing inhibitor POMC. These neuropeptides in turn, in- genetic variation. For instance, 295 coding variants in activate or excite MC4R, which is a hunger repressor. swine TASRs have recently been identified after analyz- Ghrelin is secreted by the stomach when it is empty and ing the low coverage whole genome sequences of 79 do- LEP is released by adipocytes as a response to high en- mesticandwildpigsfromEuropeandAsia[18]. ergy stores [6]. CCK is a hormone secreted in the duo- Motivated by their functional relevance in the pig, we denum as a response to luminal fat and protein and is a have sequenced ten porcine canonical TASRs and 191 strong inhibitor of food intake probably by decreasing AR genes in 304 pigs from multiple breeds in 16 DNA gastric emptying and stimulating the vagus nerve [8]. pools with the aim to identify coding polymorphisms LEP and ghrelin also inhibit and excite dopamine secre- and to provide a catalogue of potentially deleterious mu- tion, respectively [9]. Dopamine signalling in certain tations likely to affect the function of these genes. More- parts of the brain promotes appetite. Ghrelin secretion over, we have also detected indications that some of Clopetal.BMCGenomics (2016) 17:685 Page3of17 these variants might be associated with growth and fat- and one novel splice-site donor. These variants affect ness giving new insights in regard to the potential four TASRs with a clear over-representation in TAS1R1 phenotypic relevance of polymorphisms within these (Table 1). Three of these variants affecting TAS1R1 and genes. TAS1R3 had a predicted minor allele frequency (pMAF)≤0.01 (Tables 2 and 3). In addition, we identi- Results fied 125 non-synonymous-coding variants and one Sequencingstatistics codon-deletion, which are classified by snpEff as having We initially selected 459 kb of target genomic DNA a moderate (M) impact. Thirty-four of the non- (gDNA) covering the exons from the 12 TASRs and 201 synonymous changes were predicted to be deleterious AR genes. After genome capturing, sequencing and read (Mdel) bySIFT[19](Table1).TheremainingMvariants mapping, we successfully covered 372 kb of target were either SIFTpredicted astolerated (Mtol)ordidnot sequence with a read depth at each nucleotide position yield any prediction. Hence, we have identified 44 vari- (DP) above 1,000 in the 16 libraries as a whole. This ants (10 H and 34 Mdel) that are likely to have an im- corresponds to 81 % of the initial target size and 201 portant effect on swine TASR function. Remarkably, all genes fully or partially sequenced to a DP>1,000. The TASRs showed H or Mdel variants (Additional file 2). poorly sequenced genes include TAS2R3, ENSSSCG Finally, 81 variants were predicted to be synonymous 000000029894 (a swine ortholog of human TAS2R16), changes with no apparent impact (L) on the subjacent and 10 AR genes. The list of successfully sequenced proteins (Table 1). On average, the variants with strong genes is shown in Additional file 1. After genome cap- impact on protein sequence (H and Mdel) were pre- ture, sequencing and read mapping, 162,848,637 reads dictedtoberarerinthespeciesthanthose havingamild mapped tothetargetgDNAregions. impact(MtolandL)(Table2)according topMAF. Likewise, we sequenced 357,844 (80.1 % of the initial VariantidentificationinTASRandARgenes selection) bp of exonic sequence from AR genes at a We successfully sequenced (DP≥1,000) 14,598 bp DP>1,000, and identified 2,570 variant positions in AR (94.5 % of the initial selection) of TASR exons and iden- exons, most of which were bi-allelic SNVs. 1,092 tified 219 coding variants in TASR exons, 113 of which (42.4 %) variants were not annotated in dbSNP (52.1 %) do not have a dbSNP identifier and are thus (Additional file 2). Four of these positions were multi- considered novel (Additional file 2). Two TASR variants allelic with each allele predicted to have a different effect had the alternative allele fixed in the 16 pools and are onproteinsequence.Ofthe2,574assignedvarianteffects, thus likely to be either errors or private variants in the 25 had the alternative allele fixed in the 16 pools. From Duroc animal used to generate the reference sequence. the remaining 2,549 variants, 350 were classified as dele- We excluded them from our list of putative polymor- terious(HandMdel)andaffected123genes(Table1and phisms. Ten of the 217 remaining variants were Additional file 2). As for TASRs, H and Mdel as a whole classified by snpEff to have a high impact (H) on the tended to have lower pMAF than Mtol and L (Table 2). coding sequence and consequently, on the function of Weidentified74HvariantsinARgenes,29ofwhichhad the gene according to the gene annotation in the swine pMAFs≤0.01 and are thus considered rare. These H rare genome (Table 1). These include four single nucleotide variantsaffected21genes(Table3). variants (SNVs) and six short indels, which cause three We also plotted the distribution of the AR gene vari- stop-codon gains, one stop-codon loss, five frame-shift, ants predicted to have a strong impact on protein (H Table1NumberofvariantsacrosstheTASRandARgenegroupspereachimpactclass Highimpact Moderateimpact Lowimpact Gene Splice Stopgained Stoplost Startloss Frameshift Mdel Mtol/SIFT Silent Startgained %strong Total unknown impact TotalTASRs 1 3 1 0 5 34 92 80 1 20.3% 217 TAS1R1(umami) 1 1 1 0 3 4 9 12 0 32.2% 31 TAS1R3(sweetandumami) 0 1 0 0 0 1 0 11 1 14.3% 14 TAS2Rs(bitter) 0 1 0 0 2 29 83 57 0 18.5% 172 AR 37 16 0 4 17 277 615 1,559 24 14.5% 2,549 Totalsegregating 38 19 1 4 22 311 707 1,639 25 2,766 Alternativeallelefixed 4 0 0 0 4 0 9 9 1 27 Total 42 19 1 4 26 311 716 1,648 26 2,793 Splice:variantspredictedtoaltereitherdonororacceptorsplicesites;%strongimpact:percentageofH+Mdelvariants Clopetal.BMCGenomics (2016) 17:685 Page4of17 Table2VariantdistributionpereffectandpAAFwithineachgenegroup Variantfrequencyclass Strongimpact Mildimpact TASR Veryrare(pMAF<0.01) 28(63.6%) 51(29.5%) Rare(pMAF=[0.010–0.019]) 6(13.6%) 28(16.2%) Common(pMAF=[0.020–0.979]) 10(22.7%) 94(54.3%) Totalnumber 44 173 pMAFmin-max(average) 0.0016–0.1910(0.028) 0.0016–0.4920(0.091) AR Veryrare(pMAF<0.01) 184(52.6%) 560(25.5%) Rare(pMAF=[0.010-0.019]) 41(11.7%) 235(10.7%) Common(pMAF=[0.020–0.979]) 125(35.7%) 1,403(63.8%) Totalnumber 350 2,198 pMAFmin-max(average) 0.0017–0.4800(0.029) 0.0016–0.4980(0.0992) Thepercentagesareforthetotalnumberofvariantswithinthegroups:TASRstrongimpact,TASRmildimpact,ARstrongimpactandARmildimpact and Mdel) and those with a mild impact (Mtol and L) observed (p-val=0.07) and the Duroc and Landrace along the protein sequence divided in 10 consecutive were at the top (12.7 %) and bottom (9.4 %) ends, re- position bins of equal amino acid length. We noted that spectively(Table4). the strong impact variants tended to be more abundant Altogether, 26 TASR variants were present in a single at the end of the protein (bin9+bin10), and that these breed at pAAF≥0.1 and might thus be breed-specific. alsotendedtohave,onaverage,largerpMAF.Thistrend Not surprisingly, the Asian pool, which involved 15 was not observed in the set of mild impact variants Chinese Meishan and 7 Vietnamese pigs, showed the (Fig. 1). highest proportion of pool-specific variants, with 12 be- ing present only in this group (Additional file 3). The Perbreedvariantdistribution wild boar also displayed several allelic particularities, We compared the ten purebred pools and observed that, with nine unique variants. Noteworthy, seven of these as expected, larger pools contained more variants (both wild boar variants, all with similar pAAF, mapped to in TASR and AR genes). Nonetheless, the two Duroc TAS2R1 (Additional file 3). Overall, five (one H and four pools together, with 45 samples, displayed less genetic Mdel) breed-specific variants were predicted to have a diversity (1,042 variants), than the Large White, Land- strong impact on protein sequence (Additional file 3). race and the Pietrain, which had a similar number of an- TheHvariant isa stop gaininTAS1R1thatispresent in imals and were above 1,300 variants each (Table 4). The 17 % of the Mangalitza genomes, respectively. We also Asian pool harboured the largest number of variants identified two variants, a synonymous (L) and a non- (1,589 variants with a pool size=22) while the Iberian synonymous tolerated (Mtol) SNVs that affect TAS2R1, pool was ranked as the most homogenous (746 and pool that whilst being present at very high frequencies or size=13). Overall, we observed that the most ancient even fixed in all breeds (pAAF≥0.5), were absent in the breeds, Mangalitza, Iberian, Majorcan Black were less Asianpool(Additionalfile3). variableinoursetofgenes(Table4).Giventhatahigher We detected 306 AR coding variants that were number of samples seem to be correlated to a higher uniquely present or absent in one breed (Additional file number of variants, we also compared the genetic vari- 3). Of these, 35 variants involving 32 genes were pre- ability using the Watterson estimate [20] at the neutrally dicted to be of functional importance (Additional file 3). evolving synonymous sites (the third nucleotide position As in TASRs, the breed specific H and Mdel variants in within each codon), which corrects the number of the AR genes were more abundant in the Asian and the mutated sites by the number of animals included in the wild boar with 20 and five unique features, respectively. analysis. This value (Bazna: 0.00118, Mangalitza: 0.00101, It was noteworthy that two of these H and Mdel vari- Majorcan Black: 0.00103, Iberian: 0.00095, Duroc: 0.00090, ants, both in the Asian pools, were close to fixation Pietrain:0.00131,Landrace:0.00130,LargeWhite:0.00130, (pAAF≥0.9).Thesevariantsmaptotheserotoninrecep- wild boar:0.00109 and Asian: 0.00182) showed that all the tor, HTR3C, and to the cytochrome P450 gene, CYP2A6 populationshadsimilarvariabilitywiththeexceptionofthe (Additional file3). Asian pool, which genetic variability doubled that of the We performed hierarchical clustering using an Un- otherbreeds. weighted Pair Group Method with Arithmetic Mean We also compared the percentage of variants that (UPGMA) based on the 2,523 (217 TASR and 2,306 AR) were H or Mdel per breed. No obvious differences were variants that were present in at least one of the ten Clopetal.BMCGenomics (2016) 17:685 Page5of17 Table3ListofrareHvariants VariantID Effect Gene pMAF chr12_15398873_C_T STOP_GAINED ACE 0.0017 chr12_62594159_A_G START_LOST ALDH3A2 0.0086 chr7_103136276_A_G START_LOST ALDH6A1 0.0017 chr7_103129739_C_G SPLICE_SITE_ACCEPTOR ALDH6A1 0.0034 chr9_110061864_G_A STOP_GAINED CD36 0.0017 chr7_28072405_A_C SPLICE_SITE_DONOR NOTCH4 0.0017 JH118674.1_43285_C_T SPLICE_SITE_DONOR GRIA1ortholog(ENSSSCG00000024560) 0.0049 chr16_66618439_A_T START_LOST GABRG2 0.0091 chr16_66618438_C_A,T START_LOST GABRG2 0.0017 chr1_64002416_A_G SPLICE_SITE_DONOR GABRR1 0.0034 GL896494.1_7864_G_T STOP_GAINED GPR179 0.0052 chr13_203412188_T_A SPLICE_SITE_ACCEPTOR GRIK1 0.0058 chr1_77441762_G_A SPLICE_SITE_ACCEPTOR GRIK2 0.0034 chr6_49600987_C_T SPLICE_SITE_DONOR GRIN2D 0.0034 chr6_49600986_A_G SPLICE_SITE_DONOR GRIN2D 0.0034 chr13_37136736_G_T SPLICE_SITE_ACCEPTOR GRM2 0.0017 chr13_37136747_T_G SPLICE_SITE_DONOR GRM2 0.0034 chr7_34927184_A_C SPLICE_SITE_DONOR GRM4 0.0017 chr7_34893335_T_G SPLICE_SITE_ACCEPTOR GRM4 0.0017 chr18_3214922_AC_A FRAME_SHIFT HTR5A_human_ortholog(ENSSSCG00000030573) 0.0030 chr18_3151609_AG_A FRAME_SHIFT HTR5A_human_ortholog(ENSSSCG00000023549) 0.0020 chr3_55922763_C_T SPLICE_SITE_ACCEPTOR NPAS2 0.0020 chr1_14768986_T_G SPLICE_SITE_ACCEPTOR OPRM1 0.0017 chr5_85218678_C_A STOP_GAINED PAHortholog(ENSSSCG00000000856) 0.0017 chr5_85218708_C_A STOP_GAINED PAHortholog(ENSSSCG00000000856) 0.0017 chr5_85218747_G_T SPLICE_SITE_DONOR PAHortholog(ENSSSCG00000000856) 0.0017 chr3_23446168_C_T STOP_GAINED SCNN1G 0.0017 chr16_85870363_G_C SPLICE_SITE_DONOR SLC6A3 0.0017 chr16_85870364_T_A SPLICE_SITE_DONOR SLC6A3 0.0017 chr6_58114941_G_A STOP_GAINED TAS1R3 0.0016 chr6_62357984_C_T STOP_GAINED TAS1R1 0.0065 chr6_62363203_G_C STOP_LOST TAS1R1 0.0065 Thevariantidentifier(ID)containsinformationonchromosome_position_referenceallele_alternativeallele breeds and with pAAF information available in all these pAAFandphenotyperelationshipsintheF groups 2 populations. One-hundred variants were excluded from We also wanted to see whether we were able to detect the purebred comparison since they were unique to the an indication of an effect of the variants on production F animals. The resulting phylogenetic dendrogram is in traits. This was investigated by comparing the pAAFs of 2 line with what has been shown in other studies (Fig. 2). two F pools from the same experimental population, 2 Briefly, the Western (European and USA) breeds cluster each belonging to one of the tails of the phenotypic dis- together and the Asian pool form a separate branch. tribution, for average daily gain and retroperitoneal fat Within the Western cluster, the Duroc is the only mem- content. Out of the 217 TASR coding variants, 97 segre- ber of a distant branch whilst the European commercial gated in the F resource, and within these, eight 2 breeds (Large White, Landrace and Pietrain) belong to displayed significantly different pAAFs in five TASRs another sub-group and the more ancient breeds Iberian, (Table 5 and Additional file 4). Remarkably, four Majorcan Black, and Mangalitza cluster together with TAS2R4 variants showed significant differences between thewildboar. the pools. Likewise, we could compare 1,280 variants in Clopetal.BMCGenomics (2016) 17:685 Page6of17 Fig.1DistributionandpMAFofstrong(H+Mdel)andmild(Mtol+L)impactvariantsalongtheproteinbodyofARgenes.Thebarplotsfor(a) thestrongimpacteffectvariantsweremadewith310prematurestop,frame-shiftandMdel.Thebarplotforthemildimpacteffectvariants(b) weremadewith2,174Mtol,in-frameindelsandsynonymousvariants.ThedotsindicatetheaveragepMAFineachofthe10positionpercentile bins.Percentilebinsdividetheproteinbodyofeachgeneintenportionsofequalsize AR genes segregating in this population and identified pigs. We could not genotype the remaining 67 pigsasall 56 significant differences (p-val≤0.05) involving 25 their DNA was used for the sequencing step. 9 % of the genes (Table 5 and Additional file 4). After correcting assays failed to either amplify or clearly differentiate the for multiple testing (p-val≤0.05/(97+1,280)=0.00003), genotype clusters, and 46 % did not show the alternative one M variant in TAS2R39 and three variants in CD36 allele in any sample. 57 variants identified in our variant remainedsignificant (Table5). calling pipeline were validated bytheTaqManassay. The 57 polymorphisms (3 TASR H, 12 AR H, 31 TASR M Selectionandgenotype-basedvalidationofasetof and 11 AR M) affect eight TASRs and 23 AR genes protein-damagingvariants (Additional file 6). All genotyped variants showed to be With the double aim to validate the subset of variants in Hardy-Weinberg Equilibrium across breeds (data not with more likely damaging impact and, to evaluate by shown). Although none of the four genotyped rare poly- genetic association, their impact in pig breeding, we de- morphisms (all H) showed a homozygote state for the signedaTaqMangenotypingassaytargeting128putative minor allele, one H SNV (chr16_45548628_G_A), polymorphisms. We originally aimed at including all the unique to the Vietnamese pigs and with a MAF slightly H variants found in the study and to fill the remaining above the rare variant threshold (MAF=0.013), was of the assay with several M variants to cover all TASRs found in the homozygous state in two animals. This and several AR genes. However, the assay had design re- variant was predicted to cause a stop codon gain in the quirements that not all the variants fulfilled. Hence, we very last amino acid of the canonical HTR1A protein. could only include nine H variants in TASR, 69 H vari- Another introduction of a stop codon antsinAR genes,33MvariantsinTASRsand17Mvar- (chr8_38038074_G_A) was found in the middle of the iants in AR genes. To validate the selected variants, we Gamma Aminobutyric Acid receptor GABRG1, and was genotyped 237 (Additional file 5) of the 304 sequenced present only in Meishan pigs showing a MAF=0.022. C lo p e t a l. B M C G e n o m ic s (2 0 1 6 ) 1 7 Table4Numberofvariantsandnumberofuniquevariantswithineachbreedandperimpactclass :6 8 5 Duroc Pietrain Largewhite Landrace Bazna Mangalitza Iberian Majorcanblack Wildboar Asian Varianteffect nr(%) nr nr(%) nr nr(%) nr nr(%) nr nr(%) nr nr(%) nr nr(%) nr nr(%) nr nr(%) nr nr(%) nr unique unique unique unique unique unique unique unique unique unique Strongimpact 133 1 144 1 140 1 124 0 99 2 87 4 92 2 90 1 123 6 163 26 (12.7) (10.9) (10.5) (9.4) (10.2) (11.3) (12.3) (10.4) (12.4) (10.2) High 44 1 46 0 49 0 33 0 28 0 29 1 30 1 30 0 35 0 48 9 (4.2) (3.5) (3.7) (2.5) (2.9) (3.8) (4.0) (3.5) (3.5) (3.0) Moderate_Deleterious 89 0 98 1 91 1 91 0 71 2 58 3 62 1 60 1 88 6 115 17 (8.5) (7.4) (6.8) (6.9) (7.3) (7.5) (8.3) (6.9) (8.9) (7.2) Mildimpact 909 5 1177 5 1194 5 1195 3 868 11 685 8 654 9 778 21 870 32 1426 195 (87.3) (89.1) (89.5) (90.6) (89.8) (88.7) (87.7) (89.6) (87.6) (89.7) 279 4 342 1 332 2 342 1 257 2 192 2 187 5 223 9 252 15 389 50 Moderate_Tolerated (26.8) (25.9) (24.9) (25.9) (26.6) (24.9) (25.1) (25.7) (25.4) (24.5) Low 630 1 835 4 862 3 853 2 611 9 493 6 467 4 555 12 618 17 1037 145 (60.5) (63.2) (64.6) (64.7) (63.2) (63.8) (62.6) (63.9) (62.2) (65.3) Total 1042 6 1321 6 1334 6 1319 3 967 13 772 12 746 11 868 22 993 38 1589 221 (0.6%)a (0.5%)a (0.4%)a (0.2%)a (1.3%)a (1.6%)a (1.5%)a (2.5%)a (3.8%)a (13.9%)a Poolsize 45 41 39 40 15 12 13 17 22 22 Thistablewasdoneusingthe2,574variantsinTASRandARgenesincludingthemulti-allelicwithdifferenteffectsandthesewiththealternativeallelefixedinallthepopulations aPercentageofvariantsthatarebreed-specific P a g e 7 o f 1 7 Clopetal.BMCGenomics (2016) 17:685 Page8of17 Fig.2Phylogeneticdendrogramwiththe10breeds.Thenumbersindicatethesupportforeachnodeaccordingto1,000bootstrapiterations We next compared the predicted and the matched chemical molecules that could be both sources of energy genotype-observed frequency of the minor allele and or threatening toxins and promote an adequate re- detected a strong correlation (r2=0.93) between the sponse. With the aim to characterize the coding genetic two (Fig. 3). variation affecting swine TASRs and the AR circuitry We also carried a genetic association analysis with the genes, we sequenced 16 gDNA pools corresponding to genotypes and phenotypic values for retroperitoneal fat 304 pigs from ten breeds and European wild boar and and daily gain in the F . 36 of the 57 variants segregated from two pools of an experimental F population with 2 2 in the F animals. The array contains two of the variants records on growth and retroperitoneal fat content. We 2 (chr2_3566521_A_C and chr18_7019623_G_A) that have mapped thousands of coding region genetic vari- showed significant pAAF differences between the F ants, hundreds of which are expected to have a strong 2 groups with the Fisher test. For daily gain, only the impact on protein sequence, some of which are breed marker chr2_3566521_A_Cwasnominallysignificant (p- specific. By comparing the pAAFs of these variants in val=0.025) (Additional file 7) but it did not reach sig- two F pools divergent for growth and fat deposition, we 2 nificance after Bonferroni correction for multiple testing also identified many genotype - phenotype relationships. (p-val≤0.05/36=0.0013). For retroperitoneal fat, two Our data provides detailed information of the gen- markers, chr9_46397500_A_G and chr18_7019387_A_T, etic variation present in TASRs and AR genes. We were nominally significant (p-val=0.028 and 0.044, also developed an assay to genotype 128 of the most respectively) but again, none reached statistical sig- functionally relevant variants which is available to nificance after Bonferroni correction (Additional file perform association studies with relevant traits in 7). Chr18_7019623_G_A, which was significant at pig populations. the pAAF comparison between both F groups, did 2 not reach significance for any of the two traits (p- Technicalconsiderations val=0.09 and 0.1 for daily gain and retroperitoneal DifferencesinDNAextractionmethodsandinaccuracies fat, respectively). with respect to quantification might have a negative im- pact onthe even distribution of sequencingreads among Discussion the genomes within a pool thereby reducing the accur- TASRs and the components of the AR circuitry are key acyofthepAAF/pMAFsasaproxyoftherealallelicfre- genes in keeping body homeostasis as they recognize quency. Nonetheless,the comparison of predicted versus Clopetal.BMCGenomics (2016) 17:685 Page9of17 Table5GeneswithvariantsshowingsignificantlydifferentpAAFbetweenF2_FandF2_L Gene Totalnumber Numberandtypeofvariantswithsignificantdifferences FisherTestvaluerange ispAAF (min-max) ADRB1 1 1synonymous 0.0467 ALDH1B1 3 3synonymous 0.0253 ALDH2 1 1non-synoymoustolerated 0.0358 ALDH3B2 7 1non-synoymousdeleterious;6synonymous 0.0003–0.0137 ALDH6A1 1 1synonymous 0.0253 CD36 3 1frame-shift;1non-synoymousdeleterious;1synonymous 1.5xE-05a-2.58xE-05a DISC1 3 3synonymous 0.0280–0.0357 TVPR1humanortholog(ENSSSCG00000017863) 5 1non-synoymousdeleterious;4synonymous 0.0111–0.0315 ALDH8A1ortholog(ENSSSCG00000023457) 1 1synonymous 0.0387 FOS 2 1non-synoymoustolerated;1synonymous 0.0284 GABRA3 1 1splice-sitedonor/acceptor 0.0253 GABRA6 1 1synonymous 0.0047 GPR179 9 2non-synoymousdeleterious;4non-synoymoustolerated; 0.0137–0.0324 3synonymous GPRC5B 1 1non-synoymousdeleterious 0.0253 GPRC5C 1 1non-synoymoustolerated 0.0258 GRM1 3 3synonymous 0.0047–0.0383 GRM8 1 1synonymous 0.0178 HTR1B 1 1synonymous 0.0253 HTR3C 1 1synonymous 0.0357 LEPR 1 1synonymous 0.0253 MCHR2 2 2synonymous 0.0115–0.0253 MTNR1B 1 1non-synoymoustolerated 0.0357 P2RX2 2 2synonymous 0.0094–0.0324 P2X7 3 3synonymous 0.0006–0.0324 SIM1 1 1synonymous 0.0115 TAS2R16ortholog(ENSSSCG00000016433) 1 1synonymous 0.0315 TAS2R39 1 1codonchange 6.16xE-08a TAS2R4 4 2non-synoymoustolerated;2synonymous 0.0002–0.0158 TAS2R41 1 1non-synoymoustolerated 0.0207 TAS2R60 1 1synonymous 0.0351 Total 64 aVariantswithFishertestvaluesremainingsignificantafterBonferronicorrectionforthemultipletestingfor1,377variants(p-val≤0.05/(1,377)) observed allele frequencies in the 57 genotyped variants depth of sequencing of 72× for each of the haploid ge- showed that pAAF/pMAFs were very good predictors nomes. This read depth allowed us to detect all variants and supply additional confidence of the quality of our present in the pools regardless of their frequency. Given results(Fig.3). that we sequenced 304 animals harbouring in total 608 To the best of our knowledge, this is the first pub- chromosome sets (2 alleles each), we were able to detect lished study from a targeted genome enrichment experi- rare variants with MAF≥0.0016. No previous high ment in swine. As this approach reduces the sequencing throughput sequencing study in pigs reached this sensi- throughput requirements, we were able to sequence the tivity to detect rare variants. We identified a large num- target sub-genome (the exome of 201 TASRs and AR ber of variant events, most mapping to exon flanking genes) in the largest number of pigs (n=304) reported regions including promoters, introns, and upstream and to date, in a single experiment. By mapping nearly 163 downstream segments. Although some of these non- million reads on the target exons, we reached an average coding variants could be functional, their regulatory Clopetal.BMCGenomics (2016) 17:685 Page10of17 Fig.3LinearregressionandcorrelationbetweenpMAFandthegenotypeobservedMAFinthe57genotypedpolymorphisms relevance is difficult to predict and was not the subject and Tamworth). 395 of the 2,766 “segregating” variants of our study, which focused mainly, on the detection of arepredicted tohaveahigh impact(Hand Mdel) onthe protein-damaging coding variation. proteinsequence of133 genes andare thus very likelyto disrupt or strongly alter their function (Table 1 and Variantidentification Additional file 2). Some of these variants showed allelic Overall, we have identified 2,793 coding variants but 27 frequency relationship with daily gain and retro- oftheseare fixed inoursampleset and arethus likelyto peritoneal fat deposition in the F resource with pheno- 2 be either errors in or private to the reference sequence. typic records (Table 5 and Additional file 4). In keeping We annotated 217 variants segregating in the 10 canon- with previous findings in human [13], we also detected ical swineTASRs in our samples. Da Silva et al. [18] de- that H and M variants tend to be more abundant and scribed 279 coding variants in their list of 21 taste and have higher pAAFs in TASRs than in non-sensory genes nutrient receptor genes after sequencing 79 pigs. How- (Tables 1 and 2) which indicates that TASRs are subject ever, Da Silva and co-authors included 8 genes that we to balancing selection. TASRs are comparable to the did not study. These are 5 likely-paralogs of canonical major histocompatibility complex genes (HLA in TASRs and 3 genes that are not canonical TASRs but humans and SLA in pigs) as both are expressed at the that are related to tasting fat (GPR41, and GPR84) and surface of cells to detect particular molecules that could amino acids (GPRC6A). On the contrary, we included 2 be hazardous for the body. Whilst HLA and SLA detect canonical TASRs (TAS2R4 and TAS2R40) not studied by antigens to promote an immune reaction, TASRs sense Da Silva et al. Moreover, 4 we classified as AR genes chemical compounds to stimulate appropriate responses (GPR40, GPR120, CASR, GRM1 and GRM4) that are (reward, acrosomal reaction, smooth muscle contraction shared in both studies as they are not canonical TASRs or immune function among others) depending on the or are receptors belonging to the glutamate pathway. cell type that is involved. Thus, a healthy animal popula- 138 of the 279 variants, 71 in canonical TASR and 67 in tion needs to be highly polymorphic in these genes in the other shared genes, are common to both studies. order to maximize its adaptability and survival to vari- The difference in the catalogue of identified variants is able environments facing multiple threads. As expected, likely to be mostly due to the fact that Da Silva et al. H and Mdel variants had on average, lower pMAFs than studied a different set of animals, some belonging to Mtol or L variants (Table 2), as the former are more breeds we did not target (Hampshire, creole, Brazilian likely to be deleterious and consequently, subject to

Description:
A genotyping array with 57 variants in 31 of these genes is now available for genotyping and start elucidating the impact of genetic variation in these
See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.