Multifactorial methods integrating haplotype and epistasis effects for genomic estimation and prediction of quantitative traits

The rapid growth in genomic selection data provides unprecedented opportunities to discover and utilize complex genetic effects for improving phenotypes, but the methodology is lacking. Epistasis effects are interaction effects, and haplotype effects may contain local high-order epistasis effects. Multifactorial methods with SNP, haplotype, and epistasis effects up to the third-order are developed to investigate the contributions of global low-order and local high-order epistasis effects to the phenotypic variance and the accuracy of genomic prediction of quantitative traits. These methods include genomic best linear unbiased prediction (GBLUP) with associated reliability for individuals with and without phenotypic observations, including a computationally efficient GBLUP method for large validation populations, and genomic restricted maximum estimation (GREML) of the variance and associated heritability using a combination of EM-REML and AI-REML iterative algorithms. These methods were developed for two models, Model-I with 10 effect types and Model-II with 13 effect types, including intra- and inter-chromosome pairwise epistasis effects that replace the pairwise epistasis effects of Model-I. GREML heritability estimate and GBLUP effect estimate for each effect of an effect type are derived, except for third-order epistasis effects. The multifactorial models evaluate each effect type based on the phenotypic values adjusted for the remaining effect types and can use more effect types than separate models of SNP, haplotype, and epistasis effects, providing a methodology capability to evaluate the contributions of complex genetic effects to the phenotypic variance and prediction accuracy and to discover and utilize complex genetic effects for improving the phenotypes of quantitative traits.

Haplotype genomic prediction of phenotypic values based on chromosome distance and gene boundaries using low-coverage sequencing in Duroc pigs

Background

Genomic selection using single nucleotide polymorphism (SNP) markers has been widely used for genetic improvement of livestock, but most current methods of genomic selection are based on SNP models. In this study, we investigated the prediction accuracies of haplotype models based on fixed chromosome distances and gene boundaries compared to those of SNP models for genomic prediction of phenotypic values. We also examined the reasons for the successes and failures of haplotype genomic prediction.

Methods

We analyzed a swine population of 3195 Duroc boars with records on eight traits: body judging score (BJS), teat number (TN), age (AGW), loin muscle area (LMA), loin muscle depth (LMD) and back fat thickness (BF) at 100 kg live weight, and average daily gain (ADG) and feed conversion rate (FCR) from 30 to100 kg live weight. Ten-fold validation was used to evaluate the prediction accuracy of each SNP model and each multi-allelic haplotype model based on 488,124 autosomal SNPs from low-coverage sequencing. Haplotype blocks were defined using fixed chromosome distances or gene boundaries.

Results

Compared to the best SNP model, the accuracy of predicting phenotypic values using a haplotype model was greater by 7.4% for BJS, 7.1% for AGW, 6.6% for ADG, 4.9% for FCR, 2.7% for LMA, 1.9% for LMD, 1.4% for BF, and 0.3% for TN. The use of gene-based haplotype blocks resulted in the best prediction accuracy for LMA, LMD, and TN. Compared to estimates of SNP additive heritability, estimates of haplotype epistasis heritability were strongly correlated with the increase in prediction accuracy by haplotype models. The increase in prediction accuracy was largest for BJS, AGW, ADG, and FCR, which also had the largest estimates of haplotype epistasis heritability, 24.4% for BJS, 14.3% for AGW, 14.5% for ADG, and 17.7% for FCR. SNP and haplotype heritability profiles across the genome identified several genes with large genetic contributions to phenotypes: NUDT3 for LMA, LMD and BF, VRTN for TN, COL5A2 for BJS, BSND for ADG, and CARTPT for FCR.

Conclusions

Haplotype prediction models improved the accuracy for genomic prediction of phenotypes in Duroc pigs. For some traits, the best prediction accuracy was obtained with haplotypes defined using gene regions, which provides evidence that functional genomic information can improve the accuracy of haplotype genomic prediction for certain traits.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12711-021-00661-y.

Haplotype Analysis of Genomic Prediction Using Structural and Functional Genomic Information for Seven Human Phenotypes

Genomic prediction using multi-allelic haplotype models improved the prediction accuracy for all seven human phenotypes, the normality transformed high density lipoproteins, low density lipoproteins, total cholesterol, triglycerides, weight, and the original height and body mass index without normality transformation. Eight SNP sets with 40,941-380,705 SNPs were evaluated. The increase in prediction accuracy due to haplotypes was 1.86-8.12%. Haplotypes using fixed chromosome distances had the best prediction accuracy for four phenotypes, fixed number of SNPs for two phenotypes, and gene-based haplotypes for high density lipoproteins and height (tied for best). Haplotypes of coding genes were more accurate than haplotypes of all autosome genes that included both coding and noncoding genes for triglycerides and weight, and nearly the same as haplotypes of all autosome genes for the other phenotypes. Haplotypes of noncoding genes (mostly lncRNAs) only improved the prediction accuracy over the SNP models for high density lipoproteins, total cholesterol, and height. ChIP-seq haplotypes had better prediction accuracy than gene-based haplotypes for total cholesterol, body mass index and low density lipoproteins. The accuracy of ChIP-seq haplotypes was most striking for low density lipoproteins, where all four haplotype models with ChIP-seq haplotypes had similarly high prediction accuracy over the best prediction model with gene-based haplotypes. Haplotype epistasis was shown to be the reason for the increased accuracy due to haplotypes. Low density lipoproteins had the largest haplotype epistasis heritability that explained 14.70% of the phenotypic variance and was 31.27% of the SNP additive heritability, and the largest increase in prediction accuracy relative to the best SNP model (8.12%). Relative to the SNP additive heritability of the same regions, noncoding genes had the highest haplotype epistasis heritability, followed by coding genes and ChIP-seq for the seven phenotypes. SNP and haplotype heritability profiles showed that the integration of SNP and haplotype additive values compensated the weakness of haplotypes in estimating SNP heritabilities for four phenotypes, whereas models with haplotype additive values fully accounted for SNP additive values for three phenotypes. These results showed that haplotype analysis can be a method to utilize functional and structural genomic information to improve the accuracy of genomic prediction.

Genome-wide association study and accuracy of genomic prediction for teat number in Duroc pigs using genotyping-by-sequencing

Background

The number of teats in pigs is related to a sow’s ability to rear piglets to weaning age. Several studies have identified genes and genomic regions that affect teat number in swine but few common results were reported. The objective of this study was to identify genetic factors that affect teat number in pigs, evaluate the accuracy of genomic prediction, and evaluate the contribution of significant genes and genomic regions to genomic broad-sense heritability and prediction accuracy using 41,108 autosomal single nucleotide polymorphisms (SNPs) from genotyping-by-sequencing on 2936 Duroc boars.

Results

Narrow-sense heritability and dominance heritability of teat number estimated by genomic restricted maximum likelihood were 0.365 ± 0.030 and 0.035 ± 0.019, respectively. The accuracy of genomic predictions, calculated as the average correlation between the genomic best linear unbiased prediction and phenotype in a tenfold validation study, was 0.437 ± 0.064 for the model with additive and dominance effects and 0.435 ± 0.064 for the model with additive effects only. Genome-wide association studies (GWAS) using three methods of analysis identified 85 significant SNP effects for teat number on chromosomes 1, 6, 7, 10, 11, 12 and 14. The region between 102.9 and 106.0 Mb on chromosome 7, which was reported in several studies, had the most significant SNP effects in or near the PTGR2, FAM161B, LIN52, VRTN, FCF1, AREL1 and LRRC74A genes. This region accounted for 10.0% of the genomic additive heritability and 8.0% of the accuracy of prediction. The second most significant chromosome region not reported by previous GWAS was the region between 77.7 and 79.7 Mb on chromosome 11, where SNPs in the FGF14 gene had the most significant effect and accounted for 5.1% of the genomic additive heritability and 5.2% of the accuracy of prediction. The 85 significant SNPs accounted for 28.5 to 28.8% of the genomic additive heritability and 35.8 to 36.8% of the accuracy of prediction.

Conclusions

The three methods used for the GWAS identified 85 significant SNPs with additive effects on teat number, including SNPs in a previously reported chromosomal region and SNPs in novel chromosomal regions. Most significant SNPs with larger estimated effects also had larger contributions to the total genomic heritability and accuracy of prediction than other SNPs.

Electronic supplementary material

The online version of this article (doi:10.1186/s12711-017-0311-8) contains supplementary material, which is available to authorized users.