Phydelity

Phydelity identifies putative pathogen transmission clusters from phylogenetic trees using a statistically principled, phylogeny-informed, threshold-independent framework.

Key Features:

• Phylogeny-informed clustering: Uses phylogenetic relationships to define clusters rather than relying solely on pairwise genetic distances.
• Threshold-independent detection: Eliminates the need for arbitrarily defined genetic distance thresholds to determine cluster membership.
• Statistically principled framework: Identifies sequence groups that are more closely related than the ensemble distribution of the phylogeny.
• Improved cluster purity and reduced misclassification: Produces clusters with higher purity and lower misclassification probability relative to distance threshold- and model-based methods.
• No parameter calibration required: Operates without manual tuning of distance thresholds.
• Validated on simulations and empirical data: Demonstrated performance on simulated phylogenies and real datasets.
• Generalizability across pathogens: Applicable to identifying putative direct transmission events across diverse infectious diseases.

Scientific Applications:

• Transmission cluster identification: Identifying putative direct transmission events within pathogen phylogenies.
• Comparative evaluation: Benchmarking clustering quality against distance threshold- and model-based methods using simulated phylogenies.
• Pathogen surveillance: Applied to hepatitis B virus and hepatitis C virus datasets to detect clusters consistent with likely transmission links.

Methodology:

Uses a phylogeny-informed, statistically principled approach that identifies sequence groups more closely related than the ensemble distribution of the phylogeny and does not require arbitrary genetic distance thresholds.