Learned RNA outlier analysis supports rare-disease diagnosis
Source check: Source checked. Not independent scientific confirmation.
Finding: Denoising-autoencoder outlier analysis supported transcriptomic interpretation in 22 of 102 previously genetically undiagnosed patients, including 12 source-reported diagnoses. The remaining interpretations included support for known findings and proposed mechanisms. This is a clinical diagnostic-workflow yield, not 22 independently validated new diseases or a prospective comparison with standard care.
Technology: DROP; OUTRIDER; FRASER2.0; denoising autoencoders
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Autoencoders learn expected expression/splicing structure and flag patient deviations, supplying otherwise unavailable functional evidence for the scoped clinical interpretation process.
- Test and result
- The 2021–2024 clinical cohort came from two Hong Kong hospitals after routine genetic testing. Primary indexed computational Methods specify DROP1.3.3, expression/splicing outlier modules with autoencoder confounder modeling, multiple-testing correction and manual clinical review. Ten interpretations were hypothesis-driven and twelve hypothesis-free.
- Limitation / still unknown
- The authors combine model-derived outlier evidence with prior variants, phenotype and manual review; the whole diagnostic yield is not attributable to an autonomous model. This selected unsolved cohort has no randomized comparator or independent population sensitivity estimate. The six proposed mechanisms are not treated as established clinical diagnoses. Laboratory and variant-specific procedural material was not reviewed, and no operational genetic instructions are represented.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC13233572/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 21 Nov 2026. Review scope: Exact primary identity and scoped safe computational/clinical Methods and high-level Results; evidence access distinguished in case ledger; no operational laboratory procedures evaluated.
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.