FRONTEO, 大塚製薬 start AI kidney drug-target project

The project comes as AI drug discovery in Japan has moved past a boom phase into an implementation phase where companies compete on concrete results. The domestic market is growing at an average annual rate of 22.20%, and peers have already shown measurable gains: Chugai's antibody-design AI MALEXA has produced clinical-stage candidates, while Astellas integrated AI and robotics to cut the time to obtain candidate compounds by about 70% versus conventional methods.

For Otsuka, the kidney area is a long-standing strength, but it faces high unmet need in intractable diseases such as IgA nephropathy and a changing revenue structure as key product patents expire. That makes creating first-in-class drugs an urgent task, and the company is pairing its own kidney knowledge with FRONTEO's AI rather than building the capability alone.

FRONTEO's KIBIT is positioned as different from mainstream deep learning: it uses equation-based analysis of large volumes of medical literature to capture unknown relationships between diseases and molecules that papers do not directly mention. Because the learning process is lightweight enough to run on a single PC and is backed by 21 related patents, FRONTEO argues it can systematically produce discontinuous discoveries resembling human inspiration.