Genomic signatures of selection revealed by SNPs and structural variants in unselected and contemporary Holstein cattle

Authors: BOSCHIERO, CLARISSA - Council On Dairy Cattle Breeding; Rosen, Benjamin; Liu, Ge; WEBER, WANDA - University Of Minnesota; SEYKORA, ANTHONY - University Of Minnesota; Lippolis, John; CROOKER, BRIAN - University Of Minnesota; COLE, JOHN - Council On Dairy Cattle Breeding

Submitted to: World Congress of Genetics Applied in Livestock Production
Publication Type: Proceedings
Publication Acceptance Date: 4/15/2026
Publication Date: N/A
Citation: N/A

Interpretive Summary:

Technical Abstract: Since the 1960s, intensive selection in US dairy cattle has considerably increased milk yield; however, these gains have coincided with more health problems such as mastitis, metabolic disorders, and infertility. Immune system functions are particularly affected by the reduction in genomic diversity in Holsteins. Comparisons of University of Minnesota unselected Holsteins (UH) that have not been subjected to selection since 1964 and contemporary Holsteins (CH) reveal that UH cows produce less milk and exhibit better overall health than CH cows. This study aimed to detect SNPs and structural variants (SVs) from long-read sequencing data and assess selection’s impact in UH and CH cows. Thirty-four UH and 25 CH cows were sequenced on an Oxford Nanopore platform at ~30× coverage. Reads were mapped to the ARS-UCD2.0 cattle reference genome with Minimap2; and SNPs and SVs were called using Clair3 and Sniffles2, respectively. Variants were filtered for quality = 20, depth = 5, and SVs = 50 bp. Selection signatures were identified using Fixation Index (Fst) estimates from VCFtools with a 50 kb window and a step size of 20 kb. Top 1% Fst windows (SNPs) and Fst > 0.25 windows (SVs) were considered significant selection signatures. Selection signatures were further examined by overlapping them with cattle annotated genes and quantitative trait loci (QTLs). The average N50 read length for the 59 samples was ~25.6 Kb. After filtering, UH cows retained ~5.50 M SNPs and 23,196 SVs, while CH cows retained 4.74 M SNPs and 22,324 SVs. Approximately 4 M SNPs and 41,407 SVs were shared between UH and CH. From SNP data, 810 windows (top 1%) were identified as candidate selection regions, overlapping 685 genes and 3,199 QTLs from 176 traits. Genes within these regions were linked to immune functions (ITGA4, LUZP2, RCAN3, TNFAIP2, TRAF3), reproduction (AMH, PRLR, WEE2, ZP4), and milk production (IGF1R, MAP2K6, PRLR, STAT4). SV-based Fst analysis identified 2,374 (Fst > 0.25) windows under selection, encompassing 350 genes and 8,247 QTLs from 558 traits. Key genes included those linked to immunity (FCRL6, IGF2BP1, IL13, PRF1, SLAMF8), lipid metabolism (ABCA6, ACOT2, DHCR24), reproduction (GHR, POU5F1, SPATA22), and feed intake (GFRAL, HCRTR2). Both datasets revealed several olfactory receptor genes, suggesting selection related to behavioral traits. Twenty-nine genes overlapped between SNP and SV selection regions, including ATG2B, CD226, CPNE4, CSMD1, HTR4, TDRD9, and TRAPPC9. These results demonstrate that long-term artificial selection in Holsteins has shaped genomic regions associated with immunity, fertility, and metabolism. Combined SNP–SV analysis using long-read data offers new insights into the genetic basis of health and productivity under divergent selection.