Phrank is a high-performance, phenotype-driven similarity engine designed to calculate the similarity between patient cohorts. By leveraging information theory, Phrank quantifies the significance of shared phenotypic features (such as Human Phenotype Ontology terms) based on their Information Content (IC). This allows for the rarity of a shared phenotype to dictate the weight of the similarity across the cohort. The algorithm is available in Rust and Python. Check these distribution platforms for examples
Original Publication by Karthik A. Jagadeesh et al. here