Transmission clusters in the HIV-1 epidemic among men who have sex with men in Montreal, Quebec, Canada

Background. Several studies have used phylogenetics to investigate Human Immunodeficiency Virus (HIV) transmission among Men who have Sex with Men (MSMs) in Montreal, Quebec, Canada, revealing many transmission clusters. The Quebec HIV genotyping program sequence database now includes viral sequences from close to 4,000 HIV-positive individuals classified as MSMs. In this paper, we investigate clustering in those data by comparing results from several methods: the conventional Bayesian and maximum likelihood-bootstrap methods, and two more recent algorithms, DM-PhyClus, a Bayesian algorithm that produces a measure of uncertainty for proposed partitions, and the Gap Procedure, a fast distance-based approach. We estimate cluster growth by focusing on recent cases in the Primary HIV Infection (PHI) stage. Results. The analyses reveal considerable overlap between cluster estimates obtained from conventional methods. The Gap Procedure and DM-PhyClus rely on different cluster definitions and as a result, suggest moderately different partitions. All estimates lead to similar conclusions about cluster expansion: several large clusters have experienced sizeable growth, and a few new transmission clusters are likely emerging. Conclusions. The lack of a gold standard measure for clustering quality makes picking a best estimate among those proposed difficult. Work aiming to refine clustering criteria would be required to improve estimates. Nevertheless, the results unanimously stress the role that clusters play in promoting HIV incidence among MSMs.