Rare Disease Diagnosis: Connecting Phenotype, Genome & Literature

Rare disease is the hardest kind of diagnosis for a reason: by definition, few clinicians have seen the case, the signal is faint, and the answer is often buried in scattered literature. It's also where connecting evidence across domains matters most.

Three streams that rarely meet

A strong rare-disease workup pulls on three very different kinds of evidence — a structured description of the patient's observable traits (deep phenotype), the genomic signal including variants of uncertain significance, and the published literature that ties genes, variants, and phenotypes together.

• Phenotype: standardized terms describing what's actually observed.
• Genome: candidate variants, including ones no one has classified yet.
• Literature: the evidence connecting a variant or gene to a presentation.

Connection narrows the search

Individually, each stream is ambiguous. Together, they constrain each other — a phenotype makes some candidate genes far more plausible, and the literature confirms or refutes the link. Connecting them is precisely the reasoning that turns an impossible search space into a short, checkable list.

The human still decides

None of this replaces the clinician or the geneticist. It surfaces candidates with the evidence behind each, so the expert can investigate the most promising leads first. The engine does the connecting; the human makes the call.