Genome-wide association analysis for susceptibility to infection by Mycobacterium avium ssp. paratuberculosis in US Holsteins

Abstract

Paratuberculosis, or Johne's disease, is a chronic, granulomatous, gastrointestinal tract disease of cattle and other ruminants caused by the bacterium Mycobacterium avium subspecies paratuberculosis (MAP). Control of Johne's disease is based on programs of testing and culling animals positive for infection with MAP and concurrently modifying management to reduce the likelihood of infection. The current study was motivated by the hypothesis that genetic variation in host susceptibility to MAP infection can be dissected and quantifiable associations with genetic markers identified. Two separate GWAS analyses were conducted, the first using 897 genotyped Holstein artificial insemination sires with phenotypes derived from incidence of MAP infection among daughters based on milk ELISA testing records. The second GWAS analysis was a case-control design using US Holstein cows phenotyped for MAP infection by serum ELISA or fecal culture tests. Cases included cows positive for either serum ELISA, fecal culture, or both. Controls consisted of animals negative for all tests conducted. A total of 376 samples (70 cases and 306 controls) from a University of Minnesota Johne's management demonstration project and 184 samples (76 cases and 108 controls) from a Michigan State University study were used. Medium-density (sires) and high-density (cows) genotype data were imputed to full genome sequence for the analyses. Marker-trait associations were analyzed using the single-step (ss)GWAS procedure implemented in the BLUPF90 suite of programs. Evidence of significant genomic contributions for susceptibility to MAP infection were observed on multiple chromosomes. Results were combined across studies in a meta-analysis, and increased support for genomic regions on BTA7 and BTA21 were observed. Gene set enrichment analysis suggested pathways for antigen processing and presentation, antimicrobial peptides and natural killer cell-mediated cytotoxicity are relevant to variation in host susceptibility to MAP infection, among others. Genomic prediction was evaluated using a 5-fold cross-validation, and moderate correlations were observed between genomic breeding value predictions and daughter averages (∼0.43 to 0.53) for MAP infection in testing data sets. These results suggest that genomic selection against susceptibility to MAP infection is feasible in Holstein cattle.