Integrating Gene Expression Data Into Genomic Prediction.

Henner Simianer,Zhengcao Li,Johannes W R Martini,Ning Gao

doi:10.3389/fgene.2019.00126

Abstract

Gene expression profiles potentially hold valuable information for the prediction of breeding values and phenotypes. In this study, the utility of transcriptome data for phenotype prediction was tested with 185 inbred lines of Drosophila melanogaster for nine traits in two sexes. We incorporated the transcriptome data into genomic prediction via two methods: GTBLUP and GRBLUP, both combining single nucleotide polymorphisms (SNPs) and transcriptome data. The genotypic data was used to construct the common additive genomic relationship, which was used in genomic best linear unbiased prediction (GBLUP) or jointly in a linear mixed model with a transcriptome-based linear kernel (GTBLUP), or with a transcriptome-based Gaussian kernel (GRBLUP). We studied the predictive ability of the models and discuss a concept of “omics-augmented broad sense heritability” for the multi-omics era. For most traits, GRBLUP and GBLUP provided similar predictive abilities, but GRBLUP explained more of the phenotypic variance. There was only one trait (olfactory perception to Ethyl Butyrate in females) in which the predictive ability of GRBLUP (0.23) was significantly higher than the predictive ability of GBLUP (0.21). Our results suggest that accounting for transcriptome data has the potential to improve genomic predictions if transcriptome data can be included on a larger scale.

Highlights

Prediction of genetic values has been a key problem in quantitative genetics
Compared to GTBLUP, GRBLUP captured more genetic variance explained by gene expression data for some traits, especially for some traits with relatively low single nucleotide polymorphisms (SNPs)-based genomic heritability h2G, such as food intake (FI), OPMS, OPIC, OP1H, and OPEB in females and AST, olfactory perceptions to 2-Heptanone (OP2H), OPMS, and OP2H in males
Compared to GTBLUP, GRBLUP captured more genetic variance explained by gene expression data for some traits, especially for traits with relatively lower genomic heritability h2G in genomic best linear unbiased prediction (GBLUP), such as FI, OPMS, OPIC, OP1H, OPEB in female; and AST, OP2H, OPMS, OP2H in male

Summary

Introduction

Prediction of genetic values has been a key problem in quantitative genetics. Since Meuwissen et al (2001) published the landmark article, which uses whole genome single nucleotide polymorphisms (SNPs) to modify the traditional prediction of breeding values using family relationship, the concept of “genomic selection” has revolutionized animal and plant breeding. A number of statistical approaches have been applied in practice, such as genomic best linear unbiased prediction (GBLUP) (VanRaden, 2008), ridge regression (Whittaker et al, 1999), or the “Bayesian Alphabet” (Gianola et al, 2009; Gianola, 2013) These approaches utilizing genome-wide SNP data have been used to increase the genetic progress of breeding programs by increasing predictive accuracy of breeding values, reducing generation intervals or shortening the breeding cycles. For more than two genes, higher order interactions may be included, which makes the estimation of epistatic effects very difficult by using typically parametric regression methods Another problem for the prediction of phenotypes is that from DNA sequences to phenotypes there are complex biological processes that may affect the phenotypes. In the context of defining the respective broad sense heritability for the combination of genotypic data and omics data, the classical concept only covers the proportion of genetic factors including additive or dominance effects and interactions (Lush, 1940). We discuss the concept of “omics-augmented broad sense heritability” to be used in the context of the prediction of phenotypes based on effects at the genome level, and accounting for effects of downstream biological regulation captured by omics data

Methods

Results

Discussion

Conclusion