Evidence of early genomic selection in Holstein Friesian across African and European ecosystems

Abstract

Background: The Holstein Friesian (HF) cattle breed is the most dominant breed in commercial dairy farming worldwide and managed in more than 150 countries. These countries span diverse agro-climatic zones, ranging from tropical to cold regions. The introduction of HF animals in these regions occurred at different moments in the past which are poorly recorded and continued through importation of live animal and frozen semen. We hypothesize that the HF cattle populations in these regions underwent early forms of adaptation to these specific local environments. However, the detection of genetic variation associated with this adaptation remains poorly documented.

Results: This study investigates genetic relationship and potential early selection signatures in HF populations from three African countries (Egypt, South Africa, Uganda) and three European countries (Finland, Portugal, The Netherlands), considering five animals per country. Approximately 16.0 million single nucleotide polymorphisms (SNPs) were detected in the 30 HF animals and used for further analyses.

Across all countries, we identified dispersed regions totaling 3.3 megabase of ecosystem-specific genomic regions (43 genes), indicative of early selection signatures based on fixation indices (F-statistic, Fst). Furthermore, comparing variants between tropical (Egypt and Uganda) and cold regions (Finland and The Netherlands) by Fst, nucleotide diversity (θ_π ratio), and extended haplotype homozygosity (XP-EHH), we identified a total of 10 candidate regions, comprising 12 genes within a 0.57 megabase size. The regions were enriched with genes involved in signaling pathways associated directly or indirectly with adaptation, including the immune system (PGLYRP4,PGLYRP3, PAG1, CD48, SLAMF1, DYSF,and LOC615223), organ development and reproduction (LDB3, ADAMTSL4, TPRN, CCDC40, OR2AG1G, and OR8B3), thermogenic activation (TBC1D16), phospholipid metabolism (PLPPR4 and PITPNB), thermos-tolerance (ZNF423), and stimulus response (NCOA7, CYP2C85, and ARFGEF3).

Conclusion: This study provides new insights into early forms of genetic plasticity of animals adapted to very diverse ecosystems. Our findings highlight candidate genes related to immune response, organ development, reproduction, metabolism, and thermo-tolerance, hypothesizing their role in facilitating adaptation to different environments.

Supplementary Information: The online version contains supplementary material available at 10.1186/s12864-025-11828-y.

Keywords: Adaptation; Dairy farming; Holstein Friesian; Selective sweeps; Whole-genome resequencing.

Fig. 1

HF sampling and SNPs distribution from various countries. a Geographic location of 30 HF included in this study. The map images were created by authors using https://impactlab.org/map and GPS data in R package: maps. Sampling countries included Egypt (n = 5), Finland (n = 5), Portugal (n = 5), South Africa (n = 5), The Netherlands (n = 5), and Uganda (n = 5). b Chromosome-wise SNP distribution heat map across HF cattle genomes. The horizontal axis shows the chromosome length across 1 Mb windows; The pink heat map and bar plot show the chromosome density (SNP count/kb). The blue bar plot represents the calculation of gene density (genes count/Mb) across 1 Mb windows

Fig. 2

Genetic structure of HF across different countries. a PCA showing PC1 versus PC2 of 30 HF across six countries. b PCA showing PC1 versus PC2 of 30 HF and 42 previously published HF genomes. c Hierarchical clustering (Ward method) on the distance matrix based on autosomal SNP data (n = 72). d Admixture patterns of 30 HF for K = 2

Fig. 3

PCA depicting the genetic structure of potential selected alleles of SNPs between different countries

Fig. 4

Genome-wide selection signatures observed in each country. Manhattan plots depicting the Fst values (y-axis) in windows of 50 kb using a 25 kb slide across all autosomes (x-axis). Names of genes with protein-coding variants were highlighted. Using the top 0.1% of Fst values as significant outliers

Fig. 5

Selective sweeps within the HF tropical cattle. a Manhattan plots depicting the Fst values (y-axis) in windows of non-overlapped 25 kb slides across all autosomes (x-axis). Names of genes with protein-coding variants were highlighted. The red line indicates the significant threshold: top 1% for Fst (0.20). b and c Haplotype structures of the overlapped selective sweeps (candidate genes: PLPPR4: 43.92–43.98 Mb; PITPNB: 67.23–67.34 Mb) on BTA3 and BTA17. Rows correspond to individual animals, while columns represent polymorphic positions in the taurine cattle reference genome. The green color indicates presence of the reference allele, while grey represents the alternative allele. d and eθ_π ratio analyses of the most significant sweeps and surrounding regions on PLPPR4 and PITPNB genes, comparing tropical (Uganda and Egypt) and colder (Finland and The Netherlands) zones by a 25 kb sliding window. The grey color highlights the regions of candidate genes

Fig. 6

Annotation of candidate genes. a and b XP-EHH analyses of the most significant sweeps and surrounding regions on PLPPR4 and PITPNB genes, comparing tropical (Uganda and Egypt) and colder (Finland and The Netherlands) zones by a non-overlapped 25 kb sliding window. The grey color highlights the regions of candidate genes. c and d The allele frequency of alternative alleles on the candidate genes of PLPPR4 and PITPNB. e GO functional enrichment analysis for candidate genes identified within selective sweep regions. The GO analysis was divided into two categories: Biological Process (BP) and Molecular Function (MF). The point colors refer to the significance of the pathway terms (False discovery rate p-value < 0.05)