Ares Genetics to expand ARESdb proprietary contents

Ares Genetics to expand ARESdb proprietary contents

  • 15/12/2021

Diagnostics company Ares Genetics GmbH, developing NGS- and AI-powered solutions to combat AMR, has successfully completed Phase 1 of its collaboration with a leading U.S. CRO and will now enrich its ARESdb in a phase II study.

Ares Genetics is a subsidary of OpGen Inc. and aims at enabling better infection control and more targeted patient care by providing unprecedented insight into AMR through NGS and AI powered solutions. The innovative AI powered and machine learning (ML) driven approach leverages the extensive, curated database ARESdb to comprehensively detect and determine the genetic profile of pathogens.

Conventional culture-based diagnostics can determine effective antibiotics through antibiotic susceptibility testing (AST); however, they tend to be slow and insensitive. Ares Genetics is addressing this problem from a different angle, using artificial intelligence to accurately predict directly from genomic data if a pathogen is susceptible or resistant to a given antibiotic. Central to this approach are the datasets required to train predictive models, comprised of genome and phenotypic AMR data that have been generated under controlled conditions and a robust model training and testing framework.

Ares now has successfully completed Phase 1 of its collaboration with a leading U.S. CRO and reference lab, originally announced in August. During Phase 2, Ares will gain access to 1,000 proprietary genome and AST datasets, thereby strategically enriching ARESdb with proprietary data for key pathogens.

ARESdb, a database on AMR developed by Ares Genetics, has published several scientific studies on the performance of its predictive models as well as important considerations of model training. ARESdb not only comprehensively collects known genetic markers for AMR, but also harbors close to 80,000 datasets essential for the development and training of predictive models. This is up by over 40% from around 55,000 datasets a year ago.

Read more here: