From variants to networks: resolving low SNP overlap in dairy cattle meta-analysis through multi-level functional mapping

Abstract

The objective of this study was to perform a systematic review and multi-level meta-analysis of genome-wide association studies (GWAS) for functional traits in dairy cattle, focusing on calving performance and udder health. We hypothesized that despite the limited replication of individual SNPs across independent studies, functionally related variants would converge on common genes, biological pathways, and genomic regions. Following a systematic literature search and a two-stage filtering process, we extracted data from 20 independent articles covering seven traits: direct and maternal calving ease, direct and maternal stillbirth, clinical mastitis, somatic cell count, and somatic cell score. We first conducted a conventional SNP-level meta-analysis, which demonstrated minimal variant overlap across studies, preventing the detection of significant pooled effects. To resolve this low overlap, we mapped the associated variants to higher organizational levels, including genes, proteins, KEGG and Reactome pathways, and quantitative trait loci (QTL), and combined the statistical evidence using Fishers method. This multi-level functional mapping revealed significant biological convergence. At the gene level, we identified shared genes such as CCDC170 linking maternal calving ease and maternal stillbirth, while DCK, TMPRSS11F, and SLC4A4 were significant for udder health traits. Pathway-level meta-analysis highlighted major functional networks, demonstrating that mastitis-related traits were strongly associated with immune response, signal transduction, and cellular stress pathways, whereas maternal calving traits were dominated by endocrine and calcium signaling. Furthermore, our cross-trait overlap analysis identified 35 shared pathways and 30 shared QTL, reflecting a common biological architecture connecting calving difficulties and mastitis susceptibility. Notably, the innate immune system and signal transduction pathways consistently overlapped across multiple functional traits. In conclusion, conventional SNP-level meta-analysis has limited utility for integrating dairy cattle GWAS due to insufficient cross-study marker overlap. Moving from individual variants toward higher-order functional units effectively reveals reproducible genetic signals and functional convergence. The identified genes, biological pathways, and genomic regions offer valuable insights that can be leveraged as biological priors to improve genomic prediction models for dairy cattle health and reproduction.