Exploring the Landscape of Local and Distant Intra-vs. Inter-Chromosomal eQTLs Using TensorQTL and GCTA: A case study from 459 Liver Transcriptomes of Aged Laying Hens

Read full text Summarise with VIRAT
Authors: Alexandre Hubert, Fabien Degalez, Coralie Allain, Thierry Burlot, Louis Calvez, Marta Gòdia, Laëtitia Lagoutte, Bénédicte Lebez, Frédéric Lecerf, Ole Madsen, Tatiana Zerjal, Frédérique Pitel
+3 more Mathieu Emily, Sophie Allais, Sandrine Lagarrigue
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Expression quantitative trait loci (eQTL) mapping is a key approach in understanding how genetic variants influence gene expression and phenotypic variability. In this study, we compared the performance of two eQTL mapping tools: GCTA, widely used for quantitative traits, and TensorQTL, a recent method applied in human GTEx datasets. We assessed their performance in terms of eGene discovery, concordance, and computational efficiency using a large liver transcriptomic dataset from aged laying hens. Sustaining egg production in aging hens likely depends on efficient lipid synthesis, not only because yolk formation requires a large lipid supply, but also because lipids provide precursors for hormones and signaling molecules controlling ovulation and egg production. The liver is the central organ of lipid metabolism, making it a suitable tissue to investigate eQTLs influencing reproductive longevity. We investigated the genetic regulation of gene expression in the liver of 459 aged laying hens (90 weeks of age) from a single pure line, generating one of the largest eQTL datasets in livestock. Gene expression levels were derived from RNA-seq data that were inverse-normal transformed before association testing. Both GCTA and TensorQTL were used to map two types of eQTLs: i) local eQTLs defined as associations between an eGene and SNPs within ±500kb of its transcription start site (TSS), and ii) distant eQTLs defined as associations outside this region, including those on different chromosomes. To ensure robust detection, a 1Mb region was considered as an eQTL when at least five SNPs within a 100kb window exceeded a significance threshold and were in linkage disequilibrium (LD≥0.5). Bonferroni correction was applied to account for multiply testing considering 11,830 variables, and for distant eQTLs, 4,442 independent SNPs. Overall, 5,254 of the 11,830 genes expressed in the liver were regulated by at least one local eQTL across both methods. TensorQTL detected more eGenes (5,231) than GCTA (3,435), with 3,412 shared between both approaches, representing 65% and 99% of the eGenes detected by TensorQTL and GCTA respectively. This overlap for TensorQTL increased to 80% and 90% under more stringent p value thresholds (1e-9 and 1e-13), but the number of detected eGenes decreased within these ranges (4,038 and 3,054 respectively). The discrepancies likely originate from differences in model implementation and correction for population structure between the linear mixed model (GCTA) and linear regression without random effects (TensorQTL). TensorQTL’s GPU-based implementation drastically reduced computation time, identifying eQTLs in minutes instead of several days with GCTA.Beyond local associations, we identified 976 distant eQTLs with TensorQTL, most being intra-chromosomal (936) rather than inter-chromosomal (40). Around 75% of distant eQTLs were 500kb 1Mb from the eGene TSS, suggesting that most represent cis-regulatory effects rather than trans associations. These analyses are still ongoing.The liver eQTL map generated here provides a valuable resource for functional genomics in laying hens and offers insights into the genetic control of metabolic efficiency and reproductive persistence. Future work should include chromatin conformation analyses to better distinguish between distant cis-acting and trans-acting eQTLs.This project is funded by the European GEroNIMO N°101000236 and ANR ‘EFFICACE’ programs.

Fields of Study & Subjects

Genetic and phenotypic traits in livestock Genetic Mapping and Diversity in Plants and Animals Genetic Associations and Epidemiology

Publication details

Journal / source
HAL
Pages
https://plan.core-apps.com/pag33/abstract/2907311f-56dd-4348-bf69-ca2da809ea0f
Publication date
9 January 2026
Language
EN