Topological Data Analysis Identifies Copy Number Variation Subtypes of Bulls Associated with Fertility

Bull fertility is a polygenic trait of significant economic importance in dairy and beef cattle, controlled by multiple genetic variants, including copy number variations (CNVs). Conventional single-variant association analyses primarily evaluate marginal effects of individual loci, whereas CNV co-occurrence patterns require complementary approaches. This study applied topological data analysis (TDA) to identify groups of bulls based on genome-wide CNVR profiles and to assess their association with insemination success rate. A total of 88 bulls were included, and the presence/absence of 406 consensus CNVR was determined for each animal. TDA was performed using the Mapper algorithm with a two-dimensional UMAP projection as the lens and DBSCAN for clustering, followed by community detection using the Leiden method. Bootstrap resampling (100 iterations) was used to assess the stability of the identified communities, and their association with fertility was tested using ANOVA and permutation testing (n = 1000). TDA with bootstrap consensus identified five bull communities ranging from 10 to 30 individuals each. The identified subtypes differed in mean insemination success rates, ranging from 48.3% to 72.5%. These differences were assessed using ANOVA and permutation testing. Gene annotation of the loci distinguishing these communities using Ensembl release 110 for the ARS-UCD1.2 assembly revealed genes potentially involved in embryonic development, cellular signaling, and immune response (e.g., IGF2, INS, NOTCH1, PAEP, TRAF2). TDA provides a framework for identifying groups of bulls with distinct combinations of CNVRs and evaluating their potential association with fertility. Due to the sample size (n = 88) and unadjusted phenotypes, the findings should be considered exploratory and hypothesis-generating. The discovered candidate genomic regions require mandatory validation and independent replication in larger cohorts before being considered for implementation in breeding programs.