enGenome, a startup that combines AI and bioinformatics for genome analysis, has been acknowledged as a best performing predictor of diagnostic variants in the Rare Genome Project (RGP) challenge.
The challenge, organized by the NIH-funded Center for Critical Assessment of Genome Interpretation (CAGI), aims to discover state-of-the-art solutions for identifying causative variants in children with rare diseases. Thirty families were involved on behalf of the RGP, a patient-driven project led by genomics experts and clinicians at the Broad Institute of MIT and Harvard.
Participants in the challenge were provided whole-genome sequence data of solved and undiagnosed patients and their families (e.g. trio, quad) along with their clinical phenotypes.
Independent assessors evaluated participants’ predictions in providing an automated molecular diagnosis for solved patients. The best performing predictors were then applied to undiagnosed patients to identify new potentially causal variants.
enGenome participated in the RGP challenge with its variant interpretation software eVai, able to apply artificial intelligence on genomic, family and clinical data to automatically suggest the most likely molecular diagnosis of the patient.
eVai from enGenome, resulted as a best performing predictor among worldwide top academic and industrial solutions and was able to successfully diagnose two unsolved patients, increasing the diagnostic yield by 12.5%.
eVai was the only predictor capable of revealing a challenging disease-causing combination consisting of a coding and a deep-intronic variant for one of the undiagnosed patients, a newborn with a neurodevelopmental disorder.
Susanna Zucca, CSO and co-founder of enGenome, presented the method at Berkely University and considers this recognition an important mileston
