Evaluation of sequencing strategies for whole-genome imputation with hybrid peeling

The size and composition of haplotype reference panels impact the accuracy of imputation from low-pass sequencing in cattle

BACKGROUND

Low-pass sequencing followed by sequence variant genotype imputation is an alternative to the routine microarray-based genotyping in cattle. However, the impact of haplotype reference panels and their interplay with the coverage of low-pass whole-genome sequencing data have not been sufficiently explored in typical livestock settings where only a small number of reference samples is available.

METHODS

Sequence variant genotyping accuracy was compared between two variant callers, GATK and DeepVariant, in 50 Brown Swiss cattle with sequencing coverages ranging from 4- to 63-fold. Haplotype reference panels of varying sizes and composition were built with DeepVariant based on 501 individuals from nine breeds. High-coverage sequence data for 24 Brown Swiss cattle were downsampled to between 0.01- and 4-fold to mimic low-pass sequencing. GLIMPSE was used to infer sequence variant genotypes from the low-pass sequencing data using different haplotype reference panels. The accuracy of the sequence variant genotypes that were inferred from low-pass sequencing data was compared with sequence variant genotypes called from high-coverage data.

RESULTS

DeepVariant was used to establish bovine haplotype reference panels because it outperformed GATK in all evaluations. Within-breed haplotype reference panels were more accurate and efficient to impute sequence variant genotypes from low-pass sequencing than equally-sized multibreed haplotype reference panels for all target sample coverages and allele frequencies. F1 scores greater than 0.9, which indicate high harmonic means of recall and precision of called genotypes, were achieved with 0.25-fold sequencing coverage when large breed-specific haplotype reference panels (n = 150) were used. In absence of such large within-breed haplotype panels, variant genotyping accuracy from low-pass sequencing could be increased either by adding non-related samples to the haplotype reference panel or by increasing the coverage of the low-pass sequencing data. Sequence variant genotyping from low-pass sequencing was substantially less accurate when the reference panel lacked individuals from the target breed.

CONCLUSIONS

Variant genotyping is more accurate with DeepVariant than GATK. DeepVariant is therefore suitable to establish bovine haplotype reference panels. Medium-sized breed-specific haplotype reference panels and large multibreed haplotype reference panels enable accurate imputation of low-pass sequencing data in a typical cattle breed.

Genomic prediction with whole-genome sequence data in intensely selected pig lines

Background

Early simulations indicated that whole-genome sequence data (WGS) could improve the accuracy of genomic predictions within and across breeds. However, empirical results have been ambiguous so far. Large datasets that capture most of the genomic diversity in a population must be assembled so that allele substitution effects are estimated with high accuracy. The objectives of this study were to use a large pig dataset from seven intensely selected lines to assess the benefits of using WGS for genomic prediction compared to using commercial marker arrays and to identify scenarios in which WGS provides the largest advantage.

Methods

We sequenced 6931 individuals from seven commercial pig lines with different numerical sizes. Genotypes of 32.8 million variants were imputed for 396,100 individuals (17,224 to 104,661 per line). We used BayesR to perform genomic prediction for eight complex traits. Genomic predictions were performed using either data from a standard marker array or variants preselected from WGS based on association tests.

Results

The accuracies of genomic predictions based on preselected WGS variants were not robust across traits and lines and the improvements in prediction accuracy that we achieved so far with WGS compared to standard marker arrays were generally small. The most favourable results for WGS were obtained when the largest training sets were available and standard marker arrays were augmented with preselected variants with statistically significant associations to the trait. With this method and training sets of around 80k individuals, the accuracy of within-line genomic predictions was on average improved by 0.025. With multi-line training sets, improvements of 0.04 compared to marker arrays could be expected.

Conclusions

Our results showed that WGS has limited potential to improve the accuracy of genomic predictions compared to marker arrays in intensely selected pig lines. Thus, although we expect that larger improvements in accuracy from the use of WGS are possible with a combination of larger training sets and optimised pipelines for generating and analysing such datasets, the use of WGS in the current implementations of genomic prediction should be carefully evaluated against the cost of large-scale WGS data on a case-by-case basis.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12711-022-00756-0.

Rare and population-specific functional variation across pig lines

BACKGROUND

It is expected that functional, mainly missense and loss-of-function (LOF), and regulatory variants are responsible for most phenotypic differences between breeds and genetic lines of livestock species that have undergone diverse selection histories. However, there is still limited knowledge about the existing missense and LOF variation in commercial livestock populations, in particular regarding population-specific variation and how it can affect applications such as across-breed genomic prediction.

METHODS

We re-sequenced the whole genome of 7848 individuals from nine commercial pig lines (average sequencing coverage: 4.1×) and imputed whole-genome genotypes for 440,610 pedigree-related individuals. The called variants were categorized according to predicted functional annotation (from LOF to intergenic) and prevalence level (number of lines in which the variant segregated; from private to widespread). Variants in each category were examined in terms of their distribution along the genome, alternative allele frequency, per-site Wright's fixation index (F_ST), individual load, and association to production traits.

RESULTS

Of the 46 million called variants, 28% were private (called in only one line) and 21% were widespread (called in all nine lines). Genomic regions with a low recombination rate were enriched with private variants. Low-prevalence variants (called in one or a few lines only) were enriched for lower allele frequencies, lower F_ST, and putatively functional and regulatory roles (including LOF and deleterious missense variants). On average, individuals carried fewer private deleterious missense alleles than expected compared to alleles with other predicted consequences. Only a small subset of the low-prevalence variants had intermediate allele frequencies and explained small fractions of phenotypic variance (up to 3.2%) of production traits. The significant low-prevalence variants had higher per-site F_ST than the non-significant ones. These associated low-prevalence variants were tagged by other more widespread variants in high linkage disequilibrium, including intergenic variants.

CONCLUSIONS

Most low-prevalence variants have low minor allele frequencies and only a small subset of low-prevalence variants contributed detectable fractions of phenotypic variance of production traits. Accounting for low-prevalence variants is therefore unlikely to noticeably benefit across-breed analyses, such as the prediction of genomic breeding values in a population using reference populations of a different genetic background.

The construction of a haplotype reference panel using extremely low coverage whole genome sequences and its application in genome-wide association studies and genomic prediction in Duroc pigs

Extremely low coverage whole genome sequencing (lcWGS) is an economical technique to obtain high-density single nucleotide polymorphisms (SNPs). Here, we explored the feasibility of constructing a haplotype reference panel (lcHRP) using lcWGS and evaluated the effects of lcHRP through a genome-wide association study (GWAS) and genomic prediction in pigs. A total of 297 and 974 Duroc pigs were genotyped using lcWGS and a 50 K SNP array, respectively. We obtained 19,306,498 SNPs using lcWGS with an accuracy of 0.984. With the help of lcHRP, the accuracy of imputation from the SNP array to lcWGS was 0.922. Compared to the SNP array findings, those from the imputation-based GWAS identified more signals across four traits. With the integration of the top 1% imputation-based GWAS findings as genomic features, the accuracies of genomic prediction was improved by 6.0% to 13.2%. This study showed the great potential of lcWGS in pigs' molecular breeding.

Optimizing Sequencing Resources in Genotyped Livestock Populations Using Linear Programming

Low-cost genome-wide single-nucleotide polymorphisms (SNPs) are routinely used in animal breeding programs. Compared to SNP arrays, the use of whole-genome sequence data generated by the next-generation sequencing technologies (NGS) has great potential in livestock populations. However, sequencing a large number of animals to exploit the full potential of whole-genome sequence data is not feasible. Thus, novel strategies are required for the allocation of sequencing resources in genotyped livestock populations such that the entire population can be imputed, maximizing the efficiency of whole genome sequencing budgets. We present two applications of linear programming for the efficient allocation of sequencing resources. The first application is to identify the minimum number of animals for sequencing subject to the criterion that each haplotype in the population is contained in at least one of the animals selected for sequencing. The second application is the selection of animals whose haplotypes include the largest possible proportion of common haplotypes present in the population, assuming a limited sequencing budget. Both applications are available in an open source program LPChoose. In both applications, LPChoose has similar or better performance than some other methods suggesting that linear programming methods offer great potential for the efficient allocation of sequencing resources. The utility of these methods can be increased through the development of improved heuristics.

Characterization of a haplotype-reference panel for genotyping by low-pass sequencing in Swiss Large White pigs

BACKGROUND

The key-ancestor approach has been frequently applied to prioritize individuals for whole-genome sequencing based on their marginal genetic contribution to current populations. Using this approach, we selected 70 key ancestors from two lines of the Swiss Large White breed that have been selected divergently for fertility and fattening traits and sequenced their genomes with short paired-end reads.

RESULTS

Using pedigree records, we estimated the effective population size of the dam and sire line to 72 and 44, respectively. In order to assess sequence variation in both lines, we sequenced the genomes of 70 boars at an average coverage of 16.69-fold. The boars explained 87.95 and 95.35% of the genetic diversity of the breeding populations of the dam and sire line, respectively. Reference-guided variant discovery using the GATK revealed 26,862,369 polymorphic sites. Principal component, admixture and fixation index (F_ST) analyses indicated considerable genetic differentiation between the lines. Genomic inbreeding quantified using runs of homozygosity was higher in the sire than dam line (0.28 vs 0.26). Using two complementary approaches, we detected 51 signatures of selection. However, only six signatures of selection overlapped between both lines. We used the sequenced haplotypes of the 70 key ancestors as a reference panel to call 22,618,811 genotypes in 175 pigs that had been sequenced at very low coverage (1.11-fold) using the GLIMPSE software. The genotype concordance, non-reference sensitivity and non-reference discrepancy between thus inferred and Illumina PorcineSNP60 BeadChip-called genotypes was 97.60, 98.73 and 3.24%, respectively. The low-pass sequencing-derived genomic relationship coefficients were highly correlated (r > 0.99) with those obtained from microarray genotyping.

CONCLUSIONS

We assessed genetic diversity within and between two lines of the Swiss Large White pig breed. Our analyses revealed considerable differentiation, even though the split into two populations occurred only few generations ago. The sequenced haplotypes of the key ancestor animals enabled us to implement genotyping by low-pass sequencing which offers an intriguing cost-effective approach to increase the variant density over current array-based genotyping by more than 350-fold.

The Future of Livestock Management: A Review of Real-Time Portable Sequencing Applied to Livestock

Oxford Nanopore Technologies' MinION has proven to be a valuable tool within human and microbial genetics. Its capacity to produce long reads in real time has opened up unique applications for portable sequencing. Examples include tracking the recent African swine fever outbreak in China and providing a diagnostic tool for disease in the cassava plant in Eastern Africa. Here we review the current applications of Oxford Nanopore sequencing in livestock, then focus on proposed applications in livestock agriculture for rapid diagnostics, base modification detection, reference genome assembly and genomic prediction. In particular, we propose a future application: 'crush-side genotyping' for real-time on-farm genotyping for extensive industries such as northern Australian beef production. An initial in silico experiment to assess the feasibility of crush-side genotyping demonstrated promising results. SNPs were called from simulated Nanopore data, that included the relatively high base call error rate that is characteristic of the data, and calling parameters were varied to understand the feasibility of SNP calling at low coverages in a heterozygous population. With optimised genotype calling parameters, over 85% of the 10,000 simulated SNPs were able to be correctly called with coverages as low as 6×. These results provide preliminary evidence that Oxford Nanopore sequencing has potential to be used for real-time SNP genotyping in extensive livestock operations.

Accuracy of whole-genome sequence imputation using hybrid peeling in large pedigreed livestock populations

BACKGROUND

The coupling of appropriate sequencing strategies and imputation methods is critical for assembling large whole-genome sequence datasets from livestock populations for research and breeding. In this paper, we describe and validate the coupling of a sequencing strategy with the imputation method hybrid peeling in real animal breeding settings.

METHODS

We used data from four pig populations of different size (18,349 to 107,815 individuals) that were widely genotyped at densities between 15,000 and 75,000 markers genome-wide. Around 2% of the individuals in each population were sequenced (most of them at 1× or 2× and 37-92 individuals per population, totalling 284, at 15-30×). We imputed whole-genome sequence data with hybrid peeling. We evaluated the imputation accuracy by removing the sequence data of the 284 individuals with high coverage, using a leave-one-out design. We simulated data that mimicked the sequencing strategy used in the real populations to quantify the factors that affected the individual-wise and variant-wise imputation accuracies using regression trees.

RESULTS

Imputation accuracy was high for the majority of individuals in all four populations (median individual-wise dosage correlation: 0.97). Imputation accuracy was lower for individuals in the earliest generations of each population than for the rest, due to the lack of marker array data for themselves and their ancestors. The main factors that determined the individual-wise imputation accuracy were the genotyping status, the availability of marker array data for immediate ancestors, and the degree of connectedness to the rest of the population, but sequencing coverage of the relatives had no effect. The main factors that determined variant-wise imputation accuracy were the minor allele frequency and the number of individuals with sequencing coverage at each variant site. Results were validated with the empirical observations.

CONCLUSIONS

We demonstrate that the coupling of an appropriate sequencing strategy and hybrid peeling is a powerful strategy for generating whole-genome sequence data with high accuracy in large pedigreed populations where only a small fraction of individuals (2%) had been sequenced, mostly at low coverage. This is a critical step for the successful implementation of whole-genome sequence data for genomic prediction and fine-mapping of causal variants.