Multi-omics analysis

Genome-wide fine-mapping of transcriptome-wide association studies applicable across species

---

Abstract

Genome-wide association studies (GWAS) have played a central role in improving animal breeding programs by identifying single nucleotide polymorphisms (SNPs) associated with complex traits. However, because many of the identified SNPs reside in non-coding regions, linking them to biological mechanisms remains a challenge. Transcriptome-wide association studies (TWAS) address this limitation by integrating GWAS with gene expression data to identify gene-trait associations, improving our understanding of the molecular pathways driving complex traits. Despite the recent rapid development of TWAS methods, these methods have their shortcomings. Most methods analyze genes separately and do not consider the overlaps or cis-SNP correlations between genes. As a result, the identified genes may not be the causal genes. Other methods have been designed specifically for human populations, and thus, struggle to accommodate livestock populations with smaller effective population sizes and long-range linkage disequilibrium (LD), due to artificial selection. These factors present challenges in transferring human-specific methods to livestock species. The objectives of this study are to develop a flexible TWAS framework that does not rely on the presence of human-based LD patterns and explicitly accounts for confounding signals from non-causal SNPs and genes to improve precision. Our method considers all genes and an extensive set of SNPs, effectively accounting for confounding effects from all causal genes and variants, to address horizontal pleiotropy when estimating gene-trait associations. Additionally, our approach only relies on summary-level data, such as an LD reference panel and summary statistics from GWAS and eQTL (expression quantitative trait locus) studies. By overcoming key limitations of existing methods, our approach enables accurate causal inference across a wide range of species and enhances the efficiency of animal breeding programs. Using both simulated and real, publicly available data, we demonstrate that our TWAS approach outperforms existing TWAS methods. In simulations, our method outperforms current methods across multiple metrics, including error control, power to detect causal genes, fine-mapping resolution, and reproducibility across independent replicates.

Keywords: 2026

How to Cite:

Cheng, H., Li, D., Qu, J. & Zeng, J., (2026) “Genome-wide fine-mapping of transcriptome-wide association studies applicable across species”, World Congress on Genetics Applied to Livestock Production Digital Archive 2026(1): 2287264. doi: https://doi.org/10.31274/wcgalp.24250

Rights: 1