AI drug discovery support service utilized for the discovery of pancreatic cancer target molecule candidates - Cell proliferation inhibition effect confirmed

FRONTEO (Minato City, Tokyo Prefecture; President and CEO: Masahiro Morimoto) announced on July 23 that it has utilized its AI drug discovery support service DDAIF (Drug Discovery AI Factory) to extract pancreatic cancer drug target molecule candidates and conducted in vitro cancer cell proliferation inhibition tests, confirming certain effects. The company successfully extracted numerous highly novel target molecule candidates (17 genes) in just two days and confirmed through experiments that six of these genes inhibit pancreatic cancer cell proliferation.

"DDAIF" is an AI drug discovery support service that combines "KIBIT," an AI specialized in natural language processing (patented in Japan and the US), with the knowledge of the company's drug discovery researchers and engineers.

This service provides powerful support for researchers' decision-making in pharmaceutical development through analysis of disease-related gene networks and building hypotheses about target candidates, and has been adopted by multiple major pharmaceutical companies, building up a proven track record.

In this experiment, DDAIF was utilized to successfully extract 17 target molecule candidate genes from approximately 20,000 human genes in just two days.

In conventional drug discovery approaches, the "target discovery" process to extract such target molecules often requires more than two years. In particular, finding highly novel target molecules not mentioned in literature has been considered extremely difficult even with time investment.

Additionally, among the 17 target molecule candidate genes extracted, experiments confirmed that 6 genes inhibited pancreatic cancer cell proliferation.

Furthermore, 4 of these genes had no published papers reporting their association with pancreatic cancer, and the remaining 2 genes had only one paper report each (as of April 19, 2025), making them very highly novel target molecule candidates.

Traditional target discovery has been conducted through the approach of scrutinizing literature to find promising target candidates, and finding highly novel target molecules not mentioned in literature has been considered extremely difficult even with time investment.

These verification results prove that the company's self-developed AI "KIBIT" can discover relationships between unknown drug target molecules and diseases from known literature. The company expects that this technology, which can extract target molecule candidates in a short period, will enable dramatic acceleration of "target discovery," the most important aspect of drug discovery.

Pancreatic cancer has a 5-year survival rate of less than 10%, the lowest level by organ, and treatment options are limited, forcing reliance on highly toxic chemotherapy agents. Additionally, when chemotherapy agents become ineffective, treatment options become almost non-existent, making it a disease with extremely high unmet medical needs that requires new treatments and therapeutic methods.

When the effects of the 17 genes extracted in this experiment on pancreatic cancer cell line proliferation were examined, 6 of these genes showed approximately 40-60% cancer cell proliferation inhibition when their function was suppressed compared to no intervention. This ratio is comparable to "KRAS," known as a gene that causes pancreatic cancer and has particularly important effects on promoting cell carcinogenesis. The company assumes that the 6 genes may have high potential as pancreatic cancer drug target molecule candidates.

In the future, the company plans to work with collaborative partners such as the Institute of Science Tokyo to elucidate mechanisms of action and conduct animal experiments for the target molecule candidates that showed effects. Long-term, the company is considering licensing out to pharmaceutical companies.

Specially Appointed Professor Masayuki Murata from the Institute of Integrated Research, Institute of Science Tokyo commented: "The ability to narrow down multiple candidates for unpublished drug targets in a short period (days for some analysis subjects) is an innovative technology that was inconceivable with conventional approaches. While today's verification results are at the cellular level, I believe this is a practical example showing that KIBIT will greatly contribute to advancing, accelerating, and improving the success rate of drug discovery research. Our research group at the Institute of Science Tokyo aims for a deeper understanding of the life phenomena that form the foundation of drug discovery and cellular medicine fields by developing various cellular function analysis and evaluation methods. Going forward, through our ongoing joint research with FRONTEO, we expect to systematically advance functional verification through cellular experiments for the target molecule candidates FRONTEO has narrowed down, thereby accelerating drug discovery through identification of functionally effective molecules."