Dissecting the functional landscape of rare diseases through genomic variation in a heterogeneous cohort of 11,000 patients.

A multi-disease variant frequency database (FJD-DB) and variant/gene-disease association scores from case-control comparisons, feeding a data-driven guided reanalysis.

Aggregation of real-world genomic data from 11,084 unrelated patients across 122 diseases to build a multi-disease variant frequency database (FJD-DB, over 45 million unique variants, including ~184,000 potentially pathogenic) and variant and gene-disease association scores. Functional enrichment analyses highlighted disease-associated protein domains, pathways and phenotypes. The guided-reanalysis framework, applied to more than 1,100 unsolved inherited retinal dystrophy cases, yielded clinically relevant findings in 32 patients (3 molecular diagnoses, 25 possibly solved cases, 4 candidate genes).

Synthesis written by Geno'X. For the full original abstract, please refer to the source publication.

The strong idea is to turn genomic data already produced in routine into a reanalysis engine — precisely the point of exome/genome diagnostics, where value comes from continuous reinterpretation more than from new sequencing. The yield on solved cases remains modest (3 firm diagnoses out of 1,100) but realistic, and the approach is scalable. To follow after peer review (preprint).

Analysis by Dr Thibaut Benquey