Computational protocol: Gene Polymorphisms in African Buffalo Associated with Susceptibility to Bovine Tuberculosis Infection

Similar protocols

Protocol publication

[…] Genetic differentiation between the KNP and HiP subpopulations was tested using FST values in Genepop v4.0.10 , . Further statistical analyses were performed in Plink v1.07 (http://pngu.mgh.harvard.edu/purcell/plink/) . SNPs in this study were not removed if they were out of Hardy-Weinberg Equilibrium (HWE), as due to the social structure of a buffalo herd, non-random mating is observed and thus the expectations of HWE do not hold for this data. African buffalo display a male dominance hierarchy, with a select number of dominant bulls mating with the sexually mature females, who are often related. This suggests a degree of relatedness within a herd . The populations were analysed together in order to minimise the effect of any possible relatedness present in the individual herds. Quality control consisted of removing individuals with more than 10% missing data (genotyping failure) and SNPs that had a call rate of less than 90%. Pairwise linkage disequilibrium measures were calculated for all SNP pairs using correlations based on genotype allele counts, including those on different chromosomes. Allele frequencies, odds ratios (OR) and 95% confidence intervals were determined and the associated p-values were calculated using Fisher’s Exact test. A p-value <0.05 was regarded as significant. The Cochran-Mantel-Haenszel (CMH) test was used to obtain an average odds ratio and associated p-value for each SNP, adjusting for the two subpopulations. Genotypic associations with BTB infection were tested using the Cochran-Armitage trend test for the additive allele effect. The Cochran-Armitage trend test was considered the most applicable genotypic test for this study, as it does not assume HWE. Logistic regression analysis was also used to calculate odds ratios and 95% confidence intervals, assuming an additive effect of allele dosage. We controlled for subpopulation and sex by including those covariates in the logistic regression model. Permutation testing was performed using the max(T) permutation method with 5000 permutations, to obtain an empirical p-value for each SNP. SNPs found to be significantly associated with BTB status in the buffalo after correcting for the population structure were investigated further. Non-synonymous SNPs were submitted to the Ensembl Variant Effect Predictor in order to establish the type and position of the amino acid changes caused. […]

Pipeline specifications

Software tools Genepop, PLINK, VEP
Applications Population genetic analysis, GWAS
Organisms Bos taurus, Syncerus caffer, Homo sapiens
Diseases Mycobacterium Infections, Tuberculosis, Tuberculosis, Bovine