RESUMO
African trypanosomes of the sub-genus Trypanozoon) are eukaryotic parasitesthat cause disease in either humans or livestock. The development of genomic resources can be of great use to those interested in studying and controlling the spread of these trypanosomes. Here we present a large comparative analysis of Trypanozoon whole genomes, 83 in total, including human and animal infective African trypanosomes: 21 T. brucei brucei, 22 T. b. gambiense, 35 T. b. rhodesiense and 4 T. evansi strains, of which 21 were from Uganda. We constructed a maximum likelihood phylogeny based on 162,210 single nucleotide polymorphisms (SNPs.) The three Trypanosoma brucei sub-species and Trypanosoma evansi are not monophyletic, confirming earlier studies that indicated high similarity among Trypanosoma "sub-species". We also used discriminant analysis of principal components (DAPC) on the same set of SNPs, identifying seven genetic clusters. These clusters do not correspond well with existing taxonomic classifications, in agreement with the phylogenetic analysis. Geographic origin is reflected in both the phylogeny and clustering analysis. Finally, we used sparse linear discriminant analysis to rank SNPs by their informativeness in differentiating the strains in our data set. As few as 84 SNPs can completely distinguish the strains used in our study, and discriminant analysis was still able to detect genetic structure using as few as 10 SNPs. Our results reinforce earlier results of high genetic similarity between the African Trypanozoon. Despite this, a small subset of SNPs can be used to identify genetic markers that can be used for strain identification or other epidemiological investigations.