2.8 Phylogenetic reconstruction and principal component analysis of M. bovis from the safari park
A core-SNP matrix of M. bovis genomes from the safari was generated using a previously described algorithm (Zimpel, 2020) and subjected to ascertainment bias correction (ASC) using IQ-Tree -fconst (Nguyen, Schmidt, von Haeseler & Minh, 2015 ). The ASC-corrected matrix was used in the ModelFinder program (Kalyaanamoorthy et al., 2017) to select the best substitution model according to Bayesian Information Criterion (BIC). The best model, TVM+F+R10, was fixed to reconstruct the phylogeny using maximum likelihood (ML) algorithm with 1,000 UFBoot pseudoreplicates (Hoang et al., 2017). Mycobacterium bovisAF2122/97 was included in the analysis and five representative genomes of Mycobacterium caprae were used as outgroup (ERR1462591, ERR1462625, ERR1462617, ERR1462581).
To evaluate clustering among M. bovis isolates, a PCA (principal component analysis) was constructed based on the SNP matrix. The function princomp was used in R software 3.5.0 to generate 2- and 3-dimensions PCA graphs.