Epidemiologic Utility of a Framework for Partition Number Selection When Dissecting Hierarchically Clustered Genetic Data Evaluated on the Intestinal Parasite Cyclospora cayetanensis.

The article focuses on a novel, unbiased probabilistic framework for selecting the number of partitions when dissecting hierarchical clustering trees of parasite genotypes to inform parasitic disease outbreak investigations. This framework establishes a genetic distance threshold based on a normalized distribution of background distances to ensure that isolates grouped within partitions are statistically likely to belong to the same strain, addressing limitations of subjective partition selection in traditional hierarchical clustering. Applied to 1,137 Cyclospora cayetanensis genotypes, including 552 epidemiologically linked isolates from U.S. outbreaks, the method demonstrated 91% sensitivity and 100% specificity in correctly grouping related isolates. Compared to other statistical methods not tailored for molecular epidemiology, this framework provided superior discrimination and can be calibrated for use with other pathogens and genotyping data sets.