A rare-disease assessment becomes more informative when the clinical story and the genomic data are allowed to challenge each other. Sequencing can identify an unexpected candidate; careful examination can reveal that the candidate does not explain the person. Neither a striking resemblance nor a high computational rank is sufficient by itself to establish a molecular diagnosis.

Illustrated overview of turn observations into a question, highlighting clinical features, a working hypothesis, integrated interpretation.
Turn observations into a question: clinical features, a working hypothesis, integrated interpretation. Read the article for evidence and limitations. Original conceptual illustration.

Describe before you label

Build the phenotype from documented observations: what happened, when it began, how it progressed, and which investigations established it. Separate a measured feature from a suspected syndrome. Include findings from different organ systems, relevant exposures, and the family history. An absent feature is useful only when someone assessed it at an age when it could reasonably be expected.

The 2024 expert recommendations on innovative rare-disease diagnostics emphasize revisiting phenotype after a genomic candidate emerges. This reverse phenotyping can uncover overlooked findings, but it can also produce confirmation bias. A useful safeguard is to record which features were known before seeing the candidate and which were actively sought afterward. Read the expert recommendations.

Use ranking as a queue for review

Phenotype-driven software compares an individual’s features with established disease descriptions and combines that similarity with variant evidence. The Exomiser reanalysis study illustrates how updated computational prioritization can recover diagnoses from previously unsolved cases. It does not establish that the first-ranked result is correct for every new patient. Performance depends on the original testing, the phenotype supplied, and the available disease annotations. Read the Exomiser study.

Editorial illustration: A clinician assembling individual symptom cards with eye, heart and hand icons around a patient's story notebook, carefully connecting observations without a DNA backdrop.
Conceptual illustration accompanying this article. People and situations depicted are fictional; this is not a diagnostic image.

At case conference, separate three questions: Is the gene convincingly linked to this disorder? Does this particular variant fit the relevant mechanism and inheritance? Does the finding explain the observed presentation? Keeping these questions distinct makes disagreements easier to resolve.

Make uncertainty usable

  • State which major findings remain unexplained, rather than forcing every symptom into one syndrome.
  • Record the proposed evidence that would strengthen or weaken the hypothesis.
  • Preserve an updated phenotype summary for future reinterpretation.

Consider blended diagnoses and non-genetic explanations when the clinical pattern does not fit. The practical output is a defensible diagnostic hypothesis with explicit gaps, not simply a longer list of rare disease names.