Multi-omics analysis

Joint analysis of extensive functional annotations and multi-omics of cattle improves genomic prediction and mapping

---

Abstract

There are global efforts from multiple consortia to annotate the animal genomes. For example, the Functional Annotation of ANimal Genomes (FAANG) consortium provides annotation of the genomic regions by providing genomic, transcriptomic and epigenomic profiling. The Farm Animal Genotype-Tissue Expression (FarmGTEx) consortium provides estimates of the impact of variants on the transcriptome. Also, the Animal QTL Database (AQD) collects published data that can be used to annotate quantitative trait loci (QTL). Besides, many research labs generate their own molecular data for genome annotation. While all these efforts are generating a growing amount of data, how to use them in genomic prediction and mapping of causal variants remains unresolved. In this study, we present a new framework, called Functional-And-Evolutionary Multi-trait Importance (FAEMI) analysis. This method first applies a multi-trait Bayesian mixture model (BayesR and BayesRC) to jointly analyse variant effects of 16 million sequence variants on 16 traits in 103K cows, where Posterior Inclusion Probability (PIP) was estimated for each variant. Then, based on extensive functional annotations collected from FAANG, FarmGTEx, AQD, and many other resources, including data from the human genome and conserved sites across up to 200 species, we analyse these functional annotations against PIPs to derive a predictive score for these variants' potential to affect cattle traits. In validations using ~8000 bull data and additional heat-tolerance cow data, we show that FAEMI analysis provides a robust evaluation of the importance of functional annotation, which can guide future annotation experiments. As a result, we provide the FAEMI score for each of the 16 million variants analysed, which can serve as a new resource for functional annotation and prior information for genomic prediction. We demonstrate the applications of the FAEMI score by applying it to additional datasets, including heat tolerance and nitrogen efficiency in dairy cattle and carcass traits in beef cattle; none of these data were used in training the FAEMI score. We found that FAEMI score improved prediction accuracy by 10% and mapping of putatively causal variants for heat tolerance, identifying a candidate gene, stress-associated endoplasmic reticulum protein family member 2 (SERP2), underlying cattle heat tolerance. QTL affecting beef carcass traits are enriched significantly (p< 1e-6) in variants ranked high by FAEMI score, compared to low-FAEMI and random variants. Also, a set of variants significantly affecting dairy urea nitrogen concentration acted as eQTL and are also tagged by multiple epigenomic (ChIP-seq) marks. Our work reveals the importance of annotation and demonstrates the way of incorporating it in genomic selection.

Keywords: 2026

How to Cite:

Xiang, R., Breen, E., Bolormaa, S., van den Berg, I., Pryce, J., Chamberlain, A. & Goddard, M., (2026) “Joint analysis of extensive functional annotations and multi-omics of cattle improves genomic prediction and mapping”, World Congress on Genetics Applied to Livestock Production Digital Archive 2026(1): 2284485. doi: https://doi.org/10.31274/wcgalp.23570

Rights: 1