Summary of tests of association between bTB and single locus genotype.

The first 10 markers are first round candidates, indicated by a ‘C’ in the first column. Other markers comprise three groups flanking INRA111, BMS2753 and BMS495, indicated by numbers 1, 2 and 3 respectively in the first column. Each flanking marker is ordered around the first round association (dashed lines in table). Values in bold type are significant at P≤0.05, uncorrected for multiple tests. Heterozygosity (scored 0,1) is tested using general linear models of the form S∼G+H+G*H, where S = status (reactor/non-reactor), G = group (B = breed or G = STRUCTURE group) and H = heterozygosity. Significance was tested by using ANOVA to compare models with and without the term(s) deleted (either all heterozygosity terms = Full, or just the interaction term = Int). Single locus genotype-phenotype associations were determined using the program GEPHAST, controlling for possible substructure at the level of breed (BRD) or STRUCTURE group (GRP). The “within groups” columns refer to GEPHAST tests performed on restricted datasets comprising only cattle assigned to each of the three STRUCTURE defined groups. PCA refers to allele-specific association tests after correction for population substructure using a principal components analysis. Of a total of 146 tests, only those significant at P*At locus BM7209 two different alleles were significant, the other at P = 0.008.