Akihiro Imura

CEO and Founder at COGNANO, Inc.

Presenter

COGNANO is a startup that leverages its proprietary big data to drive innovation in drug discovery through deep learning. While traditional methods require extensive man-hours to research antibody drug seeds, COGNANO uses large-scale language models (LLMs) to quickly identify optimal candidates. The vast collection of antibodies generated in-house is digitized by sequencing single-domain antibody-coding genes, which are then computationally evaluated to determine the best candidates for each disease. In addition, COGNANO has pioneered a novel AI-driven drug discovery approach called Inverse Biomarker Exploring Technology (IBMET). This method identifies new target molecules for unmet medical needs by analyzing large-scale antibody data. COGNANO's approach enables the rapid generation of not only effective drug seeds but also new biomarkers for unmet needs simultaneously, which can be provided through licensing services. Aiming to lead the next generation of AI-driven drug discovery following AlphaFold, the company presented its research results at NeurIPS 2023, the premier machine learning conference, where it attracted significant attention(a). In COVID-19 drug discovery, we published a paper on the world's most potent neutralizing drug seeds in collaboration with Kyoto Univ, Osaka Univ, and the Univ of Tokyo(b). These seeds, designed by COGNANO, are currently undergoing the drug approval process led by Kyoto Univ(c). Our immediate goal is to file patent applications and license new cancer-featured biomarker information and drug discovery seeds at a rate of 2-3 sets per year. (a) A joint achievement by COGNANO and Google LLC on the international AI conference NeurIPS 2023: https://arxiv.org/abs/2306.03329 (b) Discovery of a new target molecule for TNBC: Patent application number 2024-106873 (c) The world's most potent drug efficacy and structure analysis in the development of a new coronavirus antibody drug: https://www.nature.com/articles/s42003-022-03630-3