Using affinity propagation for identifying subspecies among clonal organisms: lessons from M. tuberculosis

Background: Classification and naming is a key step in the analysis, understanding and adequate management of living organisms. However, where to set limits between groups can be puzzling especially in clonal organisms. Within the Mycobacterium tuberculosis complex (MTC), the etiological agent of tuberculosis (TB), experts have first identified several groups according to their pattern at repetitive sequences, especially at the CRISPR locus (spoligotyping), and to their epidemiological relevance. Most groups such as "Beijing" found good support when tested with other loci. However, other groups such as T family and T1 subfamily (belonging to the "Euro-American" lineage) correspond to non-monophyletic groups and still need to be refined. Here, we propose to use a method called Affinity Propagation that has been successfully used in image categorization to identify relevant patterns at the CRISPR locus in MTC. Results: To adequately infer the relative divergence time between strains, we used a distance method inspired by the recent evolutionary model by Reyes et al. We first confirm that this method performs better than the Jaccard index commonly used to compare spoligotype patterns. Second, we document the support of each spoligotype family among the previous classification using affinity propagation on the international spoligotyping database SpolDB4. This allowed us to propose a consensus assignation for all SpolDB4 spoligotypes. Third, we propose new signatures to subclassify the T family. Conclusion: Altogether, this study shows how the new clustering algorithm Affinity Propagation can help building or refining clonal organims classifications. It also describes well-supported families and subfamilies among M. tuberculosis complex, especially inside the modern "Euro-American" lineage.