Use of AI Platform Improves Diagnostic Rate of Rare Diseases in Japan

Rare Daily Staff

The National Center for Child Health and Development in Japan and the U.S.-based OmniTier have demonstrated improved diagnostic rates for rare diseases in a benchmark re-analysis study using the company’s CompStor Insight variant interpretation and custom database creation platform.

NCCHD achieved a diagnostics rate of approximately 42 percent compared to its previous rate of 35 percent. OmniTier’s CompStor platform features a proprietary AI classifier that more accurately identifies causal variants. It achieves this by searching over an expanded set of annotated variants derived from a larger number of standard and custom annotation databases, before applying American College of Medical Genetics guidelines and other ranking factors to prepare the clinical draft report.

The re-analysis study by the Department of Genome Medicine at NCCHD used a cohort of 1,293 proband WES datasets generated by Illumina short-read NGS sequencing. The proband set comprised 47 individuals, 67 duos, and 1179 trios.  A proband is the first person in a family to receive testing for a suspected genetic disease.

Of the 1,293 datasets, causal variants for 543 probands were successfully identified. The causal variants were identified by single nucleotide variants in 523 cases, and copy number variants in the remaining 20 cases.  Relative to its previous analysis, NCCHD achieved 42 percent, a 7 percent higher diagnostic rate, with the CompStor platform, representing successful diagnosis of 90 more probands in the cohort.

“Our platform is architected for maximal flexibility to enable our customers to use their own methodologies and custom databases along with local best practices for delivering affordable genomic healthcare,” said Hemant Thapar, CEO and founder of OmniTier.

Photo: Hemant Thapar, CEO and founder of OmniTier