Invitae Corporation announced the launch of an update to Invitae Generation with Clinical Variant Modeling, a novel machine learning approach designed to aid clinical interpretation of genetic testing results and increase the rate of definitive answers for patients. The first of its kind, developed by a multidisciplinary team of computational biologists, machine learning engineers, clinical experts and geneticists, Clinical Variant Modeling methodically leverages clinical information received at the time of testing to improve variant classification and reduce variants of uncertain significance (VUS). This new method represents the latest update to Invitae Generation, the company's platform that unifies evidence generation and systematic variant classification.

This innovation greatly reduces uncertainty for patients and increases the actionability of genetic testing. Clinical Variant Modeling was developed by leveraging Invitae's vast database of information on more than 4 million patients, including over 2 million analyzed DNA variants, and more than 100 million words of clinical descriptions. With this initial launch, it is estimated that nearly 45,000 patients will receive updated reports with more definitive answers as a result of reclassifying previous uncertain results for 11 genetic disorders associated with 17 genes.

The initial launch of Clinical Variant Modeling at Invitae specifically includes a model for genes associated with Lynch syndrome, the most common cause of hereditary colorectal cancer1. This modeling was developed from Invitae's database of more than 2 million patients for whom complete DNA sequence and copy number variant information has been generated across all known Lynch syndrome associated genes.