A Construction of Drug Design Theory based on the Hydration of Proteins

2023 Fiscal Year Final Research Report

• Project/Area Number: 21K06107
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 43040:Biophysics-related
• Research Institution: Tohoku University
• Principal Investigator: Yoshidome Takashi 東北大学, 工学研究科, 准教授 (90456830)
• Project Period (FY): 2021-04-01 – 2024-03-31
• Keywords: 水和 / 深層学習 / 積分方程式理論

Outline of Final Research Achievements

First, we completed the implementation of a deep-learning model that fast and accurate prediction of the hydration structures around proteins is possible. Our deep-learning model, referred to as the “gr Predictor”, was shared with the world. Then, by developing gr Predictor, we also implemented a deep-learning model, referred to as the “Deep GIST”, enabling us to predict the distributions of hydration-free energy around proteins. Furthermore, free-energy analysis of water molecules at the ligand-binding pocket was performed using gr Predictor and Deep GIST, and a feature of water molecules replaced with ligand upon ligand binding was elucidated.

Free Research Field

生物物理学

Academic Significance and Societal Importance of the Research Achievements

開発したgr PredictorやDeep GISTの創薬への応用が期待できる。既存のドッキングソフトは高速にリガンド結合部位を予測できる一方、予測精度が高くないことが問題である。その原因の1つとして水を顕に考慮していないことが挙げられる。gr PredictorやDeep GISTを導入することで、計算速度はそのままで予測精度が向上することが期待される。