A model‐based approach to characterize individual inbreeding at both global and local genomic scales

Inbreeding results from the mating of related individuals and may be associated with reduced fitness because it brings together deleterious variants in one individual. In general, inbreeding is estimated with respect to an arbitrary base population consisting of ancestors that are assumed unrelated. We herein propose a model‐based approach to estimate and characterize individual inbreeding at both global and local genomic scales by assuming the individual genome is a mosaic of homozygous‐by‐descent (HBD) and non‐HBD segments. The HBD segments may originate from ancestors tracing back to different periods in the past defining distinct age‐related classes. The lengths of the HBD segments are exponentially distributed with class‐specific parameters reflecting that inbreeding of older origin generates on average shorter stretches of observed homozygous markers. The model is implemented in a hidden Markov model framework that uses marker allele frequencies, genetic distances, genotyping error rates and the sequences of observed genotypes. Note that genotyping errors, low‐fold sequencing or genotype‐by‐sequencing data are easily accommodated under this framework. Based on simulations under the inference model, we show that the genomewide inbreeding coefficients and the parameters of the model are accurately estimated. In addition, when several inbreeding classes are simulated, the model captures them if their ages are sufficiently different. Complementary analyses, either on data sets simulated under more realistic models or on human, dog and sheep real data, illustrate the range of applications of the approach and how it can reveal recent demographic histories among populations (e.g., very recent bottlenecks or founder effects). The method also allows to clearly identify individuals resulting from extreme consanguineous matings.     

Comparative evaluation of genomic inbreeding parameters in seven commercial and autochthonous pig breeds

Single nucleotide polymorphism (SNP) genotyping tools, which can analyse thousands of SNPs covering the whole genome, have opened new opportunities to estimate the inbreeding level of animals directly using genome information. One of the most commonly used genomic inbreeding measures considers the proportion of the autosomal genome covered by runs of homozygosity (ROH), which are defined as continuous and uninterrupted chromosome portions showing homozygosity at all loci. In this study, we analysed the distribution of ROH in three commercial pig breeds (Italian Large White, n = 1968; Italian Duroc, n = 573; and Italian Landrace, n = 46) and four autochthonous breeds (Apulo-Calabrese, n = 90; Casertana, n = 90; Cinta Senese, n = 38; and Nero Siciliano, n = 48) raised in Italy, using SNP data generated from Illumina PorcineSNP60 BeadChip. We calculated ROH-based inbreeding coefficients (F_ROH) using ROH of different minimum length (1, 2, 4, 8, 16 Mbp) and compared them with several other genomic inbreeding coefficients (including the difference between observed and expected number of homozygous genotypes (F_HOM)) and correlated all these genomic-based measures with the pedigree inbreeding coefficient (F_PED) calculated for the pigs of some of these breeds. Autochthonous breeds had larger mean size of ROH than all three commercial breeds. F_HOM was highly correlated (0.671 to 0.985) with F_ROH measures in all breeds. Apulo-Calabrese and Casertana had the highest F_ROH values considering all ROH minimum lengths (ranging from 0.273 to 0.189 and from 0.226 to 0.152, moving from ROH of minimum size of 1 Mbp (F_ROH1) to 16 Mbp (F_ROH16)), whereas the lowest F_ROH values were for Nero Siciliano (from 0.072 to 0.051) and Italian Large White (from 0.117 to 0.042). F_ROH decreased as the minimum length of ROH increased for all breeds. Italian Duroc had the highest correlations between all F_ROH measures and F_PED (from 0.514 to 0.523) and between F_HOM and F_PED (0.485). Among all analysed breeds, Cinta Senese had the lowest correlation between F_ROH and F_PED. This might be due to the imperfect measure of F_PED, which, mainly in local breeds raised in extensive production systems, cannot consider a higher level of pedigree errors and a potential higher relatedness of the founder population. It appeared that ROH better captured inbreeding information in the analysed breeds and could complement pedigree-based inbreeding coefficients for the management of these genetic resources.     

Genome‐wide assessment of diversity and differentiation between original and modern Brown cattle populations

Identifying genomic regions involved in the differences between breeds can provide information on genes that are under the influence of both artificial and natural selection. The aim of this study was to assess the genetic diversity and differentiation among four different Brown cattle populations (two original vs. two modern populations) and to characterize the distribution of runs of homozygosity (ROH) islands using the Illumina Bovine SNP50 BeadChip genotyping data. After quality control, 34 735 SNPs and 106 animals were retained for the analyses. Larger heterogeneity was highlighted for the original populations. Patterns of genetic differentiation, multidimensional scaling, and the neighboring joining tree distinguished the modern from the original populations. The F_ST‐outlier identified several genes putatively involved in the genetic differentiation between the two groups, such as stature and growth, behavior, and adaptability to local environments. The ROH islands within both the original and the modern populations overlapped with QTL associated with relevant traits. In modern Brown (Brown Swiss and Italian Brown), ROH islands harbored candidate genes associated with milk production traits, in evident agreement with the artificial selection conducted to improve this trait in these populations. In original Brown (Original Braunvieh and Braunvieh), we identified candidate genes related with fat deposition, confirming that breeding strategies for the original Brown populations aimed to produce dual‐purpose animals. Our study highlighted the presence of several genomic regions that vary between Brown populations, in line with their different breeding histories.     

Invited review: Bioinformatic methods to discover the likely causal variant of a new autosomal recessive genetic condition using genome-wide data

In animals, new autosomal recessive genetic diseases (ARGD) arise all the time due to the regular, random mutations that occur during meiosis. In order to reduce the effect of any damaging new variant, it is necessary to find its cause. To evaluate the best way of doing this, 34 papers which found the exact location of a new genetic disease in livestock were reviewed and found to require at least two stages. In the initial stage the commonly used χ² method, applied in a case-control association analysis with single nucleotide polymorphism (SNP)-chip data, was found to have limitations and was almost always used in conjunction with a second method to locate the target region on the genome containing the variant. The commonly used methods had their drawbacks; so a new method was devised based on long runs of homozygosity, a common feature of new ARGD. This ‘autozygosity by difference’ method was found to be as good as, or better than, all the reviewed methods tested based on its ability to unambiguously find the shortest known target region in an already analysed data set. Mean target region length was found to be 4.6 megabases in the published reports. Success did not depend on the size of commercial SNP-chip used, and studies with as few as three cases and four controls were large enough to find the target region. The final stage relied on either sequencing the candidate genes found in the target region or using whole genome sequencing (WGS) on a small number of cases. Sometimes this latter method was used in conjunction with WGS on a number of control animals or resources such as the 1000 bull genomes data. Calculations showed that, in cattle, less than 15 animals would be needed in order to locate the new variant when using WGS data. This could be any combination of cases plus parents or other unrelated animals in the breed. Using WGS data, it would be necessary to search the three billion bases of the cattle genome for base positions which were homozygous for the same allele in all cases and heterozygous for that allele in parents, or not containing that homozygote in unrelated controls. This site could be confirmed on other healthy animals using much cheaper methods, and then a genetic test could be devised for that variant in order to screen the whole population and to devise a breeding programme to eliminate the disorder from the population.     

Long‐term selection for litter size in swine results in shifts in allelic frequency in regions involved in reproductive processes

High‐density genotype data were analyzed in three lines of swine that express substantial variation in sow fertility to uncover regions of the genome potentially influenced during selection for litter size traits. The experimental lines examined include the Nebraska Index Line (NIL), which has been subjected to long‐term selection for litter size; a control line derived from the same population that founded NIL; and a commercial Duroc × Hampshire (D × H) population, in which no selection for litter size was practiced. Regions of the genome potentially affected by selection for litter size traits in NIL were determined by multiple lines of evidence, including altered allelic frequency compared to the other lines, loss of heterozygosity and relative extended haplotype homozygosity. Additionally, a genome‐wide association study for litter size traits was conducted in a population based on NIL and commercial maternal line genetics. Several genomic regions identified as putative signatures of selection overlapped with QTL for litter size traits. One of these regions, located on SSC2 (13–14 Mb), includes the candidate gene P2X3R, which plays a role in implantation and sustained release of hormones associated with reproductive processes. Sequencing identified synonymous SNPs in P2X3R that are fixed in NIL but polymorphic with nearly equal frequencies in the D × H line, indicating a potential role of P2X3R in sow fertility. These results suggest that data derived from these lines can help to uncover and understand a portion of the genetic variance associated with fertility traits in swine.     

Estimations of linkage disequilibrium,effective population size and ROH‐based inbreeding coefficients in Spanish Churra sheep using imputed high‐density SNP genotypes

In this study, the availability of the Ovine HD SNP BeadChip (HD‐chip) and the development of an imputation strategy provided an opportunity to further investigate the extent of linkage disequilibrium (LD) at short distances in the genome of the Spanish Churra dairy sheep breed. A population of 1686 animals, including 16 rams and their half‐sib daughters, previously genotyped for the 50K‐chip, was imputed to the HD‐chip density based on a reference population of 335 individuals. After assessing the imputation accuracy for beagle v4.0 (0.922) and fimpute v2.2 (0.921) using a cross‐validation approach, the imputed HD‐chip genotypes obtained with beagle were used to update the estimates of LD and effective population size for the studied population. The imputed genotypes were also used to assess the degree of homozygosity by calculating runs of homozygosity and to obtain genomic‐based inbreeding coefficients. The updated LD estimations provided evidence that the extent of LD in Churra sheep is even shorter than that reported based on the 50K‐chip and is one of the shortest extents compared with other sheep breeds. Through different comparisons we have also assessed the impact of imputation on LD and effective population size estimates. The inbreeding coefficient, considering the total length of the run of homozygosity, showed an average estimate (0.0404) lower than the critical level. Overall, the improved accuracy of the updated LD estimates suggests that the HD‐chip, combined with an imputation strategy, offers a powerful tool that will increase the opportunities to identify genuine marker‐phenotype associations and to successfully implement genomic selection in Churra sheep.     

Combined approaches to reveal genes associated with litter size in Yunshang black goats

Intensive artificial selection has been imposed in Yunshang black goats, the first black specialist mutton goat breed in China, with a breeding object of improving reproductive performance, which has contributed to reshaping of the genome including the characterization of SNP, ROH and haplotype. However, variation in reproductive ability exists in the present population. A WGS was implemented in two subpopulations (polytocous group, PG, and monotocous group, MG) with evident differences of litter size. Following the mapping to reference genome, and SNP calling and pruning, three approaches – GWAS, ROH analysis and detection of signatures of selection – were employed to unveil candidate genes responsible for litter size. Consequently, 12 candidate genes containing OSBPL8 with the minimum P-value were uncovered by GWAS. Differences were observed in the pattern of ROH between two subpopulations that shared similar low inbreeding coefficients. Two ROH hotspots and 12 corresponding genes emerged from ROH pool association analysis. Based on the nS_L statistic, 15 and 61 promising genes were disclosed under selection for MG and PG respectively. Of them, some promising genes participate in ovarian function (PPP2R5C, CDC25A, ESR1, RPS26 and SERPINBs), seasonal reproduction (DIO3, BTG1 and CRYM) and metabolism (OSBPL8, SLC39A5 and SERPINBs). Our study pinpointed some novel promising genes influencing litter size, provided a comprehensive insight into genetic makeup of litter size and might facilitate selective breeding in goats.     

Comparative population genomics identified genomic regions and candidate genes associated with fruit domestication traits in peach

Crop evolution is a long‐term process involving selection by natural evolutionary forces and anthropogenic influences; however, the genetic mechanisms underlying the domestication and improvement of fruit crops have not been well studied to date. Here, we performed a population structure analysis in peach (Prunus persica) based on the genome‐wide resequencing of 418 accessions and confirmed the presence of an obvious domestication event during evolution. We identified 132 and 106 selective sweeps associated with domestication and improvement, respectively. Analysis of their tissue‐specific expression patterns indicated that the up‐regulation of selection genes during domestication occurred mostly in fruit and seeds as opposed to other organs. However, during the improvement stage, more up‐regulated selection genes were identified in leaves and seeds than in the other organs. Genome‐wide association studies (GWAS) using 4.24 million single nucleotide polymorphisms (SNPs) revealed 171 loci associated with 26 fruit domestication traits. Among these loci, three candidate genes were highly associated with fruit weight and the sorbitol and catechin content in fruit. We demonstrated that as the allele frequency of the SNPs associated with high polyphenol composition decreased during peach evolution, alleles associated with high sugar content increased significantly. This indicates that there is genetic potential for the breeding of more nutritious fruit with enhanced bioactive polyphenols without disturbing a harmonious sugar and acid balance by crossing with wild species. This study also describes the development of the genomic resources necessary for evolutionary research in peach and provides the large‐scale characterization of key agronomic traits in this crop species.     

Genome‐wide association and genome partitioning reveal novel genomic regions underlying variation in gastrointestinal nematode burden in a wild bird

Identifying the genetic architecture underlying complex phenotypes is a notoriously difficult problem that often impedes progress in understanding adaptive eco‐evolutionary processes in natural populations. Host–parasite interactions are fundamentally important drivers of evolutionary processes, but a lack of understanding of the genes involved in the host's response to chronic parasite insult makes it particularly difficult to understand the mechanisms of host life history trade‐offs and the adaptive dynamics involved. Here, we examine the genetic basis of gastrointestinal nematode (Trichostrongylus tenuis) burden in 695 red grouse (Lagopus lagopus scotica) individuals genotyped at 384 genome‐wide SNPs. We first use genome‐wide association to identify individual SNPs associated with nematode burden. We then partition genome‐wide heritability to identify chromosomes with greater heritability than expected from gene content, due to harbouring a multitude of additive SNPs with individually undetectable effects. We identified five SNPs on five chromosomes that accounted for differences of up to 556 worms per bird, but together explained at best 4.9% of the phenotypic variance. These SNPs were closely linked to genes representing a range of physiological processes including the immune system, protein degradation and energy metabolism. Genome partitioning indicated genome‐wide heritability of up to 29% and three chromosomes with excess heritability of up to 4.3% (total 8.9%). These results implicate SNPs and novel genomic regions underlying nematode burden in this system and suggest that this phenotype is somewhere between being based on few large‐effect genes (oligogenic) and based on a large number of genes with small individual but large combined effects (polygenic).     

A genomic prediction model for racecourse starts in the Thoroughbred horse

Durability traits in Thoroughbred horses are heritable, economically valuable and may affect horse welfare. The aims of this study were to test the hypotheses that (i) durability traits are heritable and (ii) genetic data may be used to predict a horse's potential to have a racecourse start. Heritability for the phenotype ‘number of 2‐ and 3‐year‐old starts’ was estimated to be  = 0.11 ± 0.02 (n = 4499). A genome‐wide association study identified SNP contributions to the trait. The neurotrimin (NTM), opioid‐binding protein/cell adhesion molecule like (OPCML) and prolylcarboxypeptidase (PRCP) genes were identified as candidate genes associated with the trait. NTM functions in brain development and has been shown to have been selected during the domestication of the horse. PRCP is an established expression quantitative trait locus involved in the interaction between voluntary exercise and body composition in mice. We hypothesise that variation at these loci contributes to the motivation of the horse to exercise, which may influence its response to the demands of the training and racing environment. A random forest with mixed effects (RFME) model identified a set of SNPs that contributed to 24.7% of the heritable variation in the trait. In an independent validation set (n = 528 horses), the cohort with high genetic potential for a racecourse start had significantly fewer unraced horses (16% unraced) than did low (27% unraced) potential horses and had more favourable race outcomes among those that raced. Therefore, the information from SNPs included in the model may be used to predict horses with a greater chance of a racecourse start.     

Coherent synthesis of genomic associations with phenotypes and home environments

Local adaptation is often studied via (i) multiple common garden experiments comparing performance of genotypes in different environments and (ii) sequencing genotypes from multiple locations and characterizing geographic patterns in allele frequency. Both approaches aim to characterize the same pattern (local adaptation), yet the complementary information from each has not yet been coherently integrated. Here, we develop a genome‐wide association model of genotype interactions with continuous environmental gradients (G × E), that is reaction norms. We present an approach to impute relative fitness, allowing us to coherently synthesize evidence from common garden and genome–environment associations. Our approach identifies loci exhibiting environmental clines where alleles are associated with higher fitness in home environments. Simulations show our approach can increase power to detect loci causing local adaptation. In a case study on Arabidopsis thaliana, most identified SNPs exhibited home allele advantage and fitness trade‐offs along climate gradients, suggesting selective gradients can maintain allelic clines. SNPs exhibiting G × E associations with fitness were enriched in genic regions, putative partial selective sweeps and associations with an adaptive phenotype (flowering time plasticity). We discuss extensions for situations where only adaptive phenotypes other than fitness are available. Many types of data may point towards the loci underlying G × E and local adaptation; coherent models of diverse data provide a principled basis for synthesis.     

A genomic approach to inferring kinship reveals limited intergenerational dispersal in the yellow fever mosquito

Understanding past dispersal and breeding events can provide insight into ecology and evolution and can help inform strategies for conservation and the control of pest species. However, parent–offspring dispersal can be difficult to investigate in rare species and in small pest species such as mosquitoes. Here, we develop a methodology for estimating parent–offspring dispersal from the spatial distribution of close kin, using pairwise kinship estimates derived from genome‐wide single nucleotide polymorphisms (SNPs). SNPs were scored in 162 Aedes aegypti (yellow fever mosquito) collected from eight close‐set, high‐rise apartment buildings in an area of Malaysia with high dengue incidence. We used the SNPs to reconstruct kinship groups across three orders of kinship. We transformed the geographical distances between all kin pairs within each kinship category into axial standard deviations of these distances, then decomposed these into components representing past dispersal events. From these components, we isolated the axial standard deviation of parent–offspring dispersal and estimated neighbourhood area (91 m), median parent–offspring dispersal distance (38 m) and oviposition dispersal radius within a gonotrophic cycle (25 m). We also analysed genetic structure using distance‐based redundancy analysis and linear regression, finding isolation by distance both within and between buildings and estimating neighbourhood size at 268 individuals. These findings indicate the scale required to suppress local outbreaks of arboviral disease and to target releases of modified mosquitoes for mosquito and disease control. Our methodology is readily implementable for studies of other species, including pests and species of conservation significance. [Correction added on 09 October 2020, after first online publication: 129 m corrected to 91 m; 75 m to 38 m; 36 m to 25 m.]     

A chromosome‐scale reference genome and genome‐wide genetic variations elucidate adaptation in yak

Yak is an important livestock animal for the people indigenous to the harsh, oxygen‐limited Qinghai‐Tibetan Plateau and Hindu Kush ranges of the Himalayas. The yak genome was sequenced in 2012, but its assembly was fragmented because of the inherent limitations of the Illumina sequencing technology used to analyse it. An accurate and complete reference genome is essential for the study of genetic variations in this species. Long‐read sequences are more complete than their short‐read counterparts and have been successfully applied towards high‐quality genome assembly for various species. In this study, we present a high‐quality chromosome‐scale yak genome assembly (BosGru_PB_v1.0) constructed with long‐read sequencing and chromatin interaction technologies. Compared to an existing yak genome assembly (BosGru_v2.0), BosGru_PB_v1.0 shows substantially improved chromosome sequence continuity, reduced repetitive structure ambiguity, and gene model completeness. To characterize genetic variation in yak, we generated de novo genome assemblies based on Illumina short reads for seven recognized domestic yak breeds in Tibet and Sichuan and one wild yak from Hoh Xil. We compared these eight assemblies to the BosGru_PB_v1.0 genome, obtained a comprehensive map of yak genetic diversity at the whole‐genome level, and identified several protein‐coding genes absent from the BosGru_PB_v1.0 assembly. Despite the genetic bottleneck experienced by wild yak, their diversity was nonetheless higher than that of domestic yak. Here, we identified breed‐specific sequences and genes by whole‐genome alignment, which may facilitate yak breed identification.     

Genetic contributions to precocity traits in racing Thoroughbreds

Adaptation to early training and racing (i.e. precocity), which is highly variable in racing Thoroughbreds, has implications for the selection and training of horses. We hypothesised that precocity in Thoroughbred racehorses is heritable. Age at first sprint training session (work day), age at first race and age at best race were used as phenotypes to quantify precocity. Using high‐density SNP array data, additive SNP heritability () was estimated to be 0.17, 0.14 and 0.17 for the three traits respectively. In genome‐wide association studies (GWAS) for age at first race and age at best race, a 1.98‐Mb region on equine chromosome 18 (ECA18) was identified. The most significant association was with the myostatin (MSTN) g.66493737C>T SNP (P = 5.46 × 10^?12 and P = 1.89 × 10^?14 respectively). In addition, two SNPs on ECA1 (g.37770220G>A and g.37770305T>C) within the first intron of the serotonin receptor gene HTR7 were significantly associated with age at first race and age at best race. Although no significant associations were identified for age at first work day, the MSTN:g.66493737C>T SNP was among the top 20 SNPs in the GWAS (P = 3.98 × 10^?5). Here we have identified variants with potential roles in early adaptation to training. Although there was an overlap in genes associated with precocity and distance aptitude (i.e. MSTN), the HTR7 variants were more strongly associated with precocity than with distance. Because HTR7 is closely related to the HTR1A gene, previously implicated in tractability in young Thoroughbreds, this suggests that behavioural traits may influence precocity.     

Structural diversity and dynamics of genomic replication origins in Schizosaccharomyces pombe

DNA replication origins (ORI) in Schizosaccharomyces pombe colocalize with adenine and thymine (A+T)‐rich regions, and earlier analyses have established a size from 0.5 to over 3 kb for a DNA fragment to drive replication in plasmid assays. We have asked what are the requirements for ORI function in the chromosomal context. By designing artificial ORIs, we have found that A+T‐rich fragments as short as 100 bp without homology to S. pombe DNA are able to initiate replication in the genome. On the other hand, functional dissection of endogenous ORIs has revealed that some of them span a few kilobases and include several modules that may be as short as 25–30 contiguous A+Ts capable of initiating replication from ectopic chromosome positions. The search for elements with these characteristics across the genome has uncovered an earlier unnoticed class of low‐efficiency ORIs that fire late during S phase. These results indicate that ORI specification and dynamics varies widely in S. pombe, ranging from very short elements to large regions reminiscent of replication initiation zones in mammals.     

Monitoring age‐related trends in genomic diversity of Australian lungfish

An important challenge for conservation science is to detect declines in intraspecific diversity so that management action can be guided towards populations or species at risk. The lifespan of Australian lungfish (Neoceratodus forsteri) exceeds 80 years, and human impacts on breeding habitat over the last half century may have impeded recruitment, leaving populations dominated by old postreproductive individuals, potentially resulting in a small and declining breeding population. Here, we conduct a “single‐sample” evaluation of genetic erosion within contemporary populations of the Australian lungfish. Genetic erosion is a temporal decline in intraspecific diversity due to factors such as reduced population size and inbreeding. We examined whether young individuals showed signs of reduced genetic diversity and/or inbreeding using a novel bomb radiocarbon dating method to age lungfish nonlethally, based on ¹⁴C ratios of scales. A total of 15,201 single nucleotide polymorphic (SNP) loci were genotyped in 92 individuals ranging in age from 2 to 77 years old. Standardized individual heterozygosity and individual inbreeding coefficients varied widely within and between riverine populations, but neither was associated with age, so perceived problems with recruitment have not translated into genetic erosion that could be considered a proximate threat to lungfish populations. Conservation concern has surrounded Australian lungfish for over a century. However, our results suggest that long‐lived threatened species can maintain stable levels of intraspecific variability when sufficient reproductive opportunities exist over the course of a long lifespan.