Robust identification of regulatory variants (eQTLs) using a differential expression framework developed for RNA-sequencing

Journal of Animal Science and Biotechnology

Research

Mackenzie A. Marrella¹  Fernando H. Biase¹ 

[1] School of Animal Sciences, Virginia Polytechnic Institute and State University, Blacksburg, VA, USA;

关键词: Differential gene expression;    eQTL analysis;    Gene expression;    RNA-sequencing;    Single nucleotide polymorphism;   

DOI  :  10.1186/s40104-023-00861-0

 received in 2022-11-17, accepted in 2023-03-05,  发布年份 2023

来源: Springer

【 摘 要 】

BackgroundA gap currently exists between genetic variants and the underlying cell and tissue biology of a trait, and expression quantitative trait loci (eQTL) studies provide important information to help close that gap. However, two concerns that arise with eQTL analyses using RNA-sequencing data are normalization of data across samples and the data not following a normal distribution. Multiple pipelines have been suggested to address this. For instance, the most recent analysis of the human and farm Genotype-Tissue Expression (GTEx) project proposes using trimmed means of M-values (TMM) to normalize the data followed by an inverse normal transformation.ResultsIn this study, we reasoned that eQTL analysis could be carried out using the same framework used for differential gene expression (DGE), which uses a negative binomial model, a statistical test feasible for count data. Using the GTEx framework, we identified 35 significant eQTLs (P < 5 × 10–8) following the ANOVA model and 39 significant eQTLs (P < 5 × 10–8) following the additive model. Using a differential gene expression framework, we identified 930 and six significant eQTLs (P < 5 × 10–8) following an analytical framework equivalent to the ANOVA and additive model, respectively. When we compared the two approaches, there was no overlap of significant eQTLs between the two frameworks. Because we defined specific contrasts, we identified trans eQTLs that more closely resembled what we expect from genetic variants showing complete dominance between alleles. Yet, these were not identified by the GTEx framework.ConclusionsOur results show that transforming RNA-sequencing data to fit a normal distribution prior to eQTL analysis is not required when the DGE framework is employed. Our proposed approach detected biologically relevant variants that otherwise would not have been identified due to data transformation to fit a normal distribution.

【 授权许可】

CC BY   
© The Author(s) 2023

【 图 表 】

Fig. 3

12888_2023_4816_Article_IEq3.gif

12888_2023_4816_Article_IEq1.gif

41116_2023_36_Article_IEq71.gif

Fig. 1

41116_2023_36_Article_IEq70.gif

Fig. 3

41116_2023_36_Article_IEq58.gif

41116_2023_36_Article_IEq54.gif

Fig. 1

41116_2023_36_Article_IEq17.gif

40517_2023_256_Article_IEq142.gif