Multi-generational imputation of single nucleotide polymorphism marker genotypes and accuracy of genomic selection

Availability of high-density single nucleotide polymorphism (SNP) genotyping platforms provided unprecedented opportunities to enhance breeding programmes in livestock, poultry and plant species, and to better understand the genetic basis of complex traits. Using this genomic information, genomic breeding values (GEBVs), which are more accurate than conventional breeding values. The superiority of genomic selection is possible only when high-density SNP panels are used to track genes and QTLs affecting the trait. Unfortunately, even with the continuous decrease in genotyping costs, only a small fraction of the population has been genotyped with these high-density panels. It is often the case that a larger portion of the population is genotyped with low-density and low-cost SNP panels and then imputed to a higher density. Accuracy of SNP genotype imputation tends to be high when minimum requirements are met. Nevertheless, a certain rate of genotype imputation errors is unavoidable. Thus, it is reasonable to assume that the accuracy of GEBVs will be affected by imputation errors; especially, their cumulative effects over time. To evaluate the impact of multi-generational selection on the accuracy of SNP genotypes imputation and the reliability of resulting GEBVs, a simulation was carried out under varying updating of the reference population, distance between the reference and testing sets, and the approach used for the estimation of GEBVs. Using fixed reference populations, imputation accuracy decayed by about 0.5% per generation. In fact, after 25 generations, the accuracy was only 7% lower than the first generation. When the reference population was updated by either 1% or 5% of the top animals in the previous generations, decay of imputation accuracy was substantially reduced. These results indicate that low-density panels are useful, especially when the generational interval between reference and testing population is small. As the generational interval increases, the imputation accuracies decay, although not at an alarming rate. In absence of updating of the reference population, accuracy of GEBVs decays substantially in one or two generations at the rate of 20% to 25% per generation. When the reference population is updated by 1% or 5% every generation, the decay in accuracy was 8% to 11% after seven generations using true and imputed genotypes. These results indicate that imputed genotypes provide a viable alternative, even after several generations, as long the reference and training populations are appropriately updated to reflect the genetic change in the population.     

Evaluation of measures of correctness of genotype imputation in the context of genomic prediction: a review of livestock applications

In livestock, many studies have reported the results of imputation to 50k single nucleotide polymorphism (SNP) genotypes for animals that are genotyped with low-density SNP panels. The objective of this paper is to review different measures of correctness of imputation, and to evaluate their utility depending on the purpose of the imputed genotypes. Across studies, imputation accuracy, computed as the correlation between true and imputed genotypes, and imputation error rates, that counts the number of incorrectly imputed alleles, are commonly used measures of imputation correctness. Based on the nature of both measures and results reported in the literature, imputation accuracy appears to be a more useful measure of the correctness of imputation than imputation error rates, because imputation accuracy does not depend on minor allele frequency (MAF), whereas imputation error rate depends on MAF. Therefore imputation accuracy can be better compared across loci with different MAF. Imputation accuracy depends on the ability of identifying the correct haplotype of a SNP, but many other factors have been identified as well, including the number of genotyped immediate ancestors, the number of animals with genotypes at the high-density panel, the SNP density on the low- and high-density panel, the MAF of the imputed SNP and whether imputed SNP are located at the end of a chromosome or not. Some of these factors directly contribute to the linkage disequilibrium between imputed SNP and SNP on the low-density panel. When imputation accuracy is assessed as a predictor for the accuracy of subsequent genomic prediction, we recommend that: (1) individual-specific imputation accuracies should be used that are computed after centring and scaling both true and imputed genotypes; and (2) imputation of gene dosage is preferred over imputation of the most likely genotype, as this increases accuracy and reduces bias of the imputed genotypes and the subsequent genomic predictions.     

Accuracy of genotype imputation in Labrador Retrievers

The dog is a valuable model species for the genetic analysis of complex traits, and the use of genotype imputation in dogs will be an important tool for future studies. It is of particular interest to analyse the effect of factors like single nucleotide polymorphism (SNP) density of genotyping arrays and relatedness between dogs on imputation accuracy due to the acknowledged genetic and pedigree structure of dog breeds. In this study, we simulated different genotyping strategies based on data from 1179 Labrador Retriever dogs. The study involved 5826 SNPs on chromosome 1 representing the high density (HighD) array; the low‐density (LowD) array was simulated by masking different proportions of SNPs on the HighD array. The correlations between true and imputed genotypes for a realistic masking level of 87.5% ranged from 0.92 to 0.97, depending on the scenario used. A correlation of 0.92 was found for a likely scenario (10% of dogs genotyped using HighD, 87.5% of HighD SNPs masked in the LowD array), which indicates that genotype imputation in Labrador Retrievers can be a valuable tool to reduce experimental costs while increasing sample size. Furthermore, we show that genotype imputation can be performed successfully even without pedigree information and with low relatedness between dogs in the reference and validation sets. Based on these results, the impact of genotype imputation was evaluated in a genome‐wide association analysis and genomic prediction in Labrador Retrievers.     

MagicalRsq: Machine-learning-based genotype imputation quality calibration

1. Download : Download high-res image (200KB)
2. Download : Download full-size image

     

Runs of homozygosity reveal genome‐wide autozygosity in Italian sheep breeds

The availability of dense single nucleotide polymorphism (SNP) assays allows for the determination of autozygous segments based on runs of consecutive homozygous genotypes (ROH). The aim of the present study was to investigate the occurrence and distribution of ROH in 21 Italian sheep breeds using medium‐density SNP genotypes in order to characterize autozygosity and identify genomic regions that frequently appeared in ROH within individuals, namely ROH islands. After filtering, the final number of animals and SNPs retained for analyses were 502 and 46 277 respectively. A total of 12 302 ROH were identified. The mean number of ROH per breed ranged from 10.58 (Comisana) to 44.54 (Valle del Belice). The average length of ROH across breeds was 4.55 Mb and ranged from 3.85 Mb (Biellese) to 5.51 Mb (Leccese). Valle del Belice showed the highest value of inbreeding on the basis of ROH (F_ROH = 0.099), whereas Comisana showed the lowest (F_ROH = 0.016), and high standard deviation values revealed high variability in autozygosity levels within each breed. Differences also existed in the length of ROH. Analysis of the distribution of ROH according to their size showed that, for all breeds, the majority of the detected ROH were <10 Mb in length, with a few long ROH >25 Mb. The levels of ROH that we estimated here reflect the inbreeding history of the investigated sheep breeds. These results also highlight that ancient and recent inbreeding have had an impact on the genome of the Italian sheep breeds and suggest that several animals have experienced recent autozygosity events. Comisana and Bergamasca appeared as the less consanguineous breeds, whereas Barbaresca, Leccese and Valle del Belice showed ROH patterns typically produced by recent inbreeding. Moreover, within the genomic regions most commonly associated with ROH, several candidate genes were detected.     

Differentiation among Spanish sheep breeds using microsatellites

Genetic variability at 18 microsatellites was analysed on the basis of individual genotypes in five Spanish breeds of sheep – Churra, Latxa, Castellana, Rasa-Aragonesa and Merino -, with Awassi also being studied as a reference breed. The degree of population subdivision calculated between Spanish breeds from F_ST diversity indices was around 7% of total variability. A high degree of reliability was obtained for individual-breed assignment from the 18 loci by using different approaches among which the Bayesian method provided to be the most efficient, with an accuracy for nine microsatellites of over 99%. Analysis of the Bayesian assignment criterion illustrated the divergence between any one breed and the others, which was highest for Awassi sheep, while no great differences were evident among the Spanish breeds. Relationships between individuals were analysed from the proportion of shared alleles. The resulting dendrogram showed a remarkable breed structure, with the highest level of clustering among members of the Spanish breeds in Latxa and the lowest in Merino sheep, the latter breed exhibiting a peculiar pattern of clustering, with animals grouped into several closely set nodes. Analysis of individual genotypes provided valuable information for understanding intra- and inter-population genetic differences and allowed for a discussion with previously reported results using populations as taxonomic units.     

Genetic structure and admixture in sheep from terminal breeds in the United States

Selection for performance in diverse production settings has resulted in variation across sheep breeds worldwide. Although sheep are an important species to the United States, the current genetic relationship among many terminal sire breeds is not well characterized. Suffolk, Hampshire, Shropshire and Oxford (terminal) and Rambouillet (dual purpose) sheep (n = 248) sampled from different flocks were genotyped using the Applied Biosystems Axiom Ovine Genotyping Array (50K), and additional Shropshire sheep (n = 26) using the Illumina Ovine SNP50 BeadChip. Relationships were investigated by calculating observed heterozygosity, inbreeding coefficients, eigenvalues, pairwise Wright’s F_ST estimates and an identity by state matrix. The mean observed heterozygosity for each breed ranged from 0.30 to 0.35 and was consistent with data reported in other US and Australian sheep. Suffolk from two different regions of the United States (Midwest and West) clustered separately in eigenvalue plots and the rectangular cladogram. Further, divergence was detected between Suffolk from different regions with Wright’s F_ST estimate. Shropshire animals showed the greatest divergence from other terminal breeds in this study. Admixture between breeds was examined using admixture , and based on cross-validation estimates, the best fit number of populations (clusters) was K = 6. The greatest admixture was observed within Hampshire, Suffolk, and Shropshire breeds. When plotting eigenvalues, US terminal breeds clustered separately in comparison with sheep from other locations of the world. Understanding the genetic relationships between terminal sire breeds in sheep will inform us about the potential applicability of markers derived in one breed to other breeds based on relatedness.     

Genetic diversity and subdivision of 57 European and Middle-Eastern sheep breeds

The population structure and genetic diversity of 57 European and Middle Eastern marginal and cosmopolitan sheep breeds from 15 countries were analysed by typing 31 microsatellite markers. Mean unbiased expected heterozygosities ranged from 0.63 in British Exmoor Horn to 0.77 in Albanian Ruda. South-eastern European and Middle-Eastern sheep breeds were significantly more variable than northwestern and western European breeds. An overall heterozygote deficiency (f) across all loci was observed (P < 0.001), while genetic differentiation (theta) was 5.7%. Principal component analysis and Bayesian model-based clustering indicate a south-east to north-west cline, but also revealed distinct groups of Middle-Eastern fat-tailed sheep, south-eastern European sheep and north-western/western European sheep. Within the last group, two less-distinct clusters comprised the Merino-type and Alpine breeds respectively. The incomplete demarcations of most clusters probably reflects cross-breeding and/or upgrading.     

Resequencing and signatures of selection scan in two Siberian native sheep breeds point to candidate genetic variants for adaptation and economically important traits

Russian sheep breeds represent an important economic asset by providing meat and wool, whilst being adapted to extreme climates. By resequencing two Russian breeds from Siberia: Tuva (n = 20) and Baikal (n = 20); and comparing them with a European (UK) sheep outgroup (n = 14), 41 million variants were called, and signatures of selection were identified. High‐frequency missense mutations on top of selection peaks were found in genes related to immunity (LOC101109746) in the Baikal breed and wool traits (IDUA), cell differentiation (GLIS1) and fat deposition (AADACL3) in the Tuva breed. In addition, genes found under selection owing to haplotype frequency changes were related to wool traits (DSC2), parasite resistance (CLCA1), insulin receptor pathway (SOCS6) and DNA repair (DDB2) in the Baikal breed, and vision (GPR179) in the Tuva breed. Our results present candidate genes and SNPs for future selection programmes, which are necessary to maintain and increase socioeconomic gain from Siberian breeds.     

North European short-tailed breeds of sheep: a review

The short-tailed sheep, native of an area stretching from Russia to Iceland, are generally considered a primitive type. These robust northern sheep seem to have been spread by Norse vikings to several countries in this area from the late eighth century to the middle of the eleventh century ad. They have several common characteristics in addition to the fluke-shaped and tapered short tail, such as a wide range of colour patterns, dual-coated wool and the ability to thrive under harsh environmental conditions, often in isolated marginal areas. While 34 short-tailed breeds of North European origin can still be identified, it is clear that their population sizes have declined in most cases and several of them are now rare and endangered. Although these breeds have mainly been confined to certain localities, some of them have gained considerable distribution due to their genetic merits, such as prolificacy. Of these, the Finnsheep and the Romanov are best known being exported to several countries in the world where their genetic material has been utilized through crossbreeding with local sheep. This has resulted in the production of some new synthetic breeds. Meat is now generally the main product of the North European short-tailed breeds and their crossbreds, whereas wool, skins and milk are normally regarded as byproducts, yet of considerable economic importance in some cases. Such breeds have clearly a role to play in sustainable grassland-based production systems in the future.     

Microsatellite analysis to estimate genetic relationships among five bulgarian sheep breeds

Herein, genetic relationships among five breeds of Bulgarian sheep were estimated using microsatellite markers. The total number of alleles identified was 226 at the 16 loci examined. D(A) distance values were used for phylogenetic tree construction with the UPGMA algorithm. The two Tsigai and two Maritza populations were found to be geneticallvery closely related to each other y (0.198, and 0.258 respectively). The Pleven Black Head population was distinct from the other four. These results could be useful for preserving genes in these breeds, thereby ensuring their preservation in Bulgaria.     

Use of partial least squares regression to predict single nucleotide polymorphism marker genotypes when some animals are genotyped with a low-density panel

High-density single nucleotide polymorphism (SNP) platforms are currently used in genomic selection (GS) programs to enhance the selection response. However, the genotyping of a large number of animals with high-throughput platforms is rather expensive and may represent a constraint for a large-scale implementation of GS. The use of low-density marker (LDM) platforms could overcome this problem, but different SNP chips may be required for each trait and/or breed. In this study, a strategy of imputation independent from trait and breed is proposed. A simulated population of 5865 individuals with a genome of 6000 SNP equally distributed on six chromosomes was considered. First, reference and prediction populations were generated by mimicking high- and low-density SNP platforms, respectively. Then, the partial least squares regression (PLSR) technique was applied to reconstruct the missing SNP in the low-density chip. The proportion of SNP correctly reconstructed by the PLSR method ranged from 0.78 to 0.97 when 90% and 50%, respectively, of genotypes were predicted. Moreover, data sets consisting of a mixture of actual and PLSR-predicted SNP or only actual SNP were used to predict genomic breeding values (GEBVs). Correlations between GEBV and true breeding values varied from 0.74 to 0.76, respectively. The results of the study indicate that the PLSR technique can be considered a reliable computational strategy for predicting SNP genotypes in an LDM platform with reasonable accuracy.     

Penalized classification for optimal statistical selection of markers from high-throughput genotyping: application in sheep breeds

The identification of individuals’ breed of origin has several practical applications in livestock and is useful in different biological contexts such as conservation genetics, breeding and authentication of animal products. In this paper, penalized multinomial regression was applied to identify the minimum number of single nucleotide polymorphisms (SNPs) from high-throughput genotyping data for individual assignment to dairy sheep breeds reared in Sicily. The combined use of penalized multinomial regression and stability selection reduced the number of SNPs required to 48. A final validation step on an independent population was carried out obtaining 100% correctly classified individuals. The results using independent analysis, such as admixture, F_st, principal component analysis and random forest, confirmed the ability of these methods in selecting distinctive markers. The identified SNPs may constitute a starting point for the development of a SNP based identification test as a tool for breed assignment and traceability of animal products.     

Pronounced differences in the frequency of TaqI βA-inhibin alleles between sheep breeds with different reproductive performance

A total of 419 individuals of four breeds with differing fecundity (Rhoenschaf, Merinolandschaf, East Friesian Milksheep, Romanov) and several wild sheep of the Ovis musimon, O. orientalis, O. vignei and O. ammon groups were screened for genetic variation at the βA-inhibin (INHBA) locus with up to 11 enzymes. The four breeds differed significantly (P< 0.001) in TaqI allele frequencies. The frequency of the TaqI A allele coincided with the average litter size in each breed.     

Genetic diversity and differentiation of 12 eastern Adriatic and western Dinaric native sheep breeds using microsatellites

Nuclear genetic diversity and differentiation of 341 sheep belonging to 12 sheep breeds from Croatia and Bosnia and Herzegovina were examined. The aim of the study was to provide the understanding of the genetic structure and variability of the analysed pramenka sheep populations, and to give indications for conservation strategies based on the population diversity and structure information. The genetic variation of the sheep populations, examined at the nuclear level using 27 microsatellite loci, revealed considerable levels of genetic diversity, similar to the diversity found in other European indigenous low-production sheep breeds. Population-specific alleles were detected at most loci and in breeds analysed. The observed heterozygosity ranged from 0.643 (in Lika pramenka) to 0.743 (in Vlasic pramenka), and the expected heterozygosity ranged from 0.646 (in Lika pramenka) to 0.756 (in Dalmatian pramenka). Significant inbreeding coefficients were found for half of the populations studied and ranged from 0.040 (Pag island sheep) to 0.091 (Kupres pramenka). Moderate genetic differentiation was found between the studied sheep populations. The total genetic variability observed between different populations was 5.29%, whereas 94.71% of the variation was found within populations. Cres island sheep, Lika pramenka and Istrian sheep were identified as the most distinct populations, which was confirmed by the factorial analysis of correspondence and supported through a bootstrapping adjustment to correct for the difference in the sample sizes. The population structure analysis distinguished 12 clusters for the 12 sheep breeds analysed. However, the cluster differentiation was low for Dalmatian, Vlasic, Stolac and Krk pramenka. This systematic study identified Lika pramenka and Rab island sheep as those with the lowest diversity, whereas Istrian sheep and Pag island sheep had the highest. Conservation actions are proposed for Istrian, Rab and Cres island sheep, Lika and Kupres pramenka because of high estimated coefficients of inbreeding.     

Genetic diversity in nine native Turkish sheep breeds based on microsatellite analysis

The genetic variability of 204 sheep from nine breeds (Awassi, Çine Çapar?, Karaka?, Karya, Karayaka, Morkaraman, Norduz, Sak?z and Tuj) growing in four different regions (western, northern, eastern and south‐eastern Anatolia) was assessed using 18 microsatellite loci. High mean number of alleles, allelic richness and factorial correspondence analysis showed the degree of admixture between native sheep breeds of Turkey. The Karya and Çine Çapar? breeds were observed as the most distinct of the breeds, and possible introgressions were detected in other breeds. It was found that 8.9% of genetic variation resulted from the difference between the populations. The genetic variation in Turkish breeds was not much higher than that of European breeds, which might be a consequence of the recent sharp decrease in sheep numbers.     

Prospects and limits of marker imputation in quantitative genetic studies in European elite wheat (Triticum aestivum L.)

Background

The main goal of our study was to investigate the implementation, prospects, and limits of marker imputation for quantitative genetic studies contrasting map-independent and map-dependent algorithms. We used a diversity panel consisting of 372 European elite wheat (Triticum aestivum L.) varieties, which had been genotyped with SNP arrays, and performed intensive simulation studies.

Results

Our results clearly showed that imputation accuracy was substantially higher for map-dependent compared to map-independent methods. The accuracy of marker imputation depended strongly on the linkage disequilibrium between the markers in the reference panel and the markers to be imputed. For the decay of linkage disequilibrium present in European wheat, we concluded that around 45,000 markers are needed for low cost, low-density marker profiling. This will facilitate high imputation accuracy, also for rare alleles. Genomic selection and diversity studies profited only marginally from imputing missing values. In contrast, the power of association mapping increased substantially when missing values were imputed.

Conclusions

Imputing missing values is especially of interest for an economic implementation of association mapping in breeding populations.

Electronic supplementary material

The online version of this article (doi:10.1186/s12864-015-1366-y) contains supplementary material, which is available to authorized users.     

Performance and utilization of Northern European short-tailed breeds of sheep and their crosses in North America: a review

The five Northern European short-tailed sheep breeds present in North America are the Finnsheep, Romanov, Icelandic, Shetland and Gotland. The Finnsheep and Romanov were first imported in 1966 and 1986, respectively, for their high reproductive performance. The Shetland, Icelandic and Gotland breeds were first imported in 1980, 1985 and 2005, respectively, for the uniqueness of their physical appearance and their unique fleeces desired by fiber craftspeople. There have been no scientific studies conducted on the performance of the Shetland, Icelandic or Gotland breeds relative to other breeds of sheep in North America. However, the Shetland and Icelandic breeds have become very popular in the United States and ranked 9th and 18th, respectively, among 35 breeds of sheep for number of purebred animals registered in 2008. The performance of the Finnsheep breed in North America relative to domestic breeds has been thoroughly investigated. Compared to several domestic purebreds and crosses, sheep with Finnsheep breeding had a younger age at puberty, greater fertility to autumn mating, greater litter size, greater survival to weaning, similar growth rate, similar subcutaneous fat thickness, smaller loin muscle area and greater percentage of kidney and pelvic fat. Each 1% increase in Finnsheep breeding in ewes was associated with approximately 0.01 more lambs born per ewe lambing. In North American studies, Romanov ewes were superior to Finnsheep ewes for reproductive rate and lamb production per ewe under both autumn and spring mating. Lambs of the two breeds were similar for survival, growth and carcass traits. Romanov and Romanov-cross ewes produced fleeces that were heavily contaminated with medulated and colored fibers and were of very low commercial value. Three composite breeds containing 25% to 49% Finnsheep breeding (Polypay, Rideau Arcott and Outaouais Arcott) were developed in North America and are now more popular than the Finnsheep breed.     

Genetic diversity of two native sheep breeds by genome-wide analysis of single nucleotide polymorphisms

Wallachian and Sumava sheep are autochthonous breeds that have undergone a significant bottleneck effect and subsequent restoration efforts. The first objective of this study was to evaluate the degree of genetic variability of both breeds and, therefore, the current management of the breeding. The second was to determine whether these two breeds still retain their genetic uniqueness in relation to each other and other breeds, despite regenerative interventions. Our data consisted of 48 individuals of Sumava and 37 individuals of Wallachian sheep. The comparison data contained 25 other breeds (primarily European) from the HapMap dataset generated by the International Sheep Genomics Consortium. When comparing all 27 breeds, the Czech breeds clustered with 15 other breeds and formed a single branch with them according to Nei's distances. At the same time, however, the clusters of both breeds were integral and easily distinguishable from the others when displayed with principal component analysis (PCA). Population substructure analysis did not show any common genetic ancestry of the Czech national breeds and breeds used for regeneration or, eventually, breeds whose ancestral population was used for regeneration. The average values of F_ST were higher in Wallachian sheep (F_ST = 0.14) than in Sumava sheep (F_ST = 0.08). The linkage disequilibrium (LD) extension per autosome was higher in Wallachian than in Sumava sheep. Consequently, the N_e estimates five generations ago were 68 for Sumava versus 34 for Wallachian sheep. Both native Czech breeds exhibit a wide range of inbreeding based on the excess of homozygosity (F_HOM) among individuals, from ?0.04 to 0.16 in Sumava and from ?0.13 to 0.12 in Wallachian. Average inbreeding based on runs of homozygosity was 0.21 in Sumava and 0.27 in Wallachian. Most detected runs of homozygosity (ROH) were less than 5 Mb long for both breeds. ROH segments longer than 15 Mb were absent in Wallachian sheep. Concerning putative selection signatures, a total of 471 candidate genes in Wallachian sheep within 11 hotspots and 653 genes within 13 hotspots in Sumava sheep were identified. Czech breeds appear to be well differentiated from each other and other European breeds. Their genetic diversity is low, especially in the case of the Wallachian breed. Sumava is not so threatened by low diversity but has a larger share of the non-native gene pool.