NLP for Clinical Phenotyping

Genotype is only half the equation. To diagnose a patient, we need their phenotype. But this data is often locked in unstructured clinical notes.

From Notes to HPO Terms

Natural Language Processing (NLP) models can parse doctor's notes and automatically extract Human Phenotype Ontology (HPO) terms. For example, converting "patient has a webbed neck" to HP:0000465.

ClinPhen and Beyond

Tools like ClinPhen and modern LLM-based extractors allow for automated prioritization of variants based on the patient's specific symptoms, drastically reducing the time to diagnosis.