Medical Molecular Informatics
Our laboratory aims to generate clinically applicable insights through computational approaches to medical data. By applying cutting-edge technologies, we address long-standing challenges in medicine and health, including the prediction of biological activities of pharmaceuticals and chemicals based on chemical structure information, the elucidation of adverse outcome pathways using artificial intelligence, and the identification of factors associated with adverse drug reactions through multivariate analysis and data mining.
Members
Research Topics
-
Prediction of Biological Activity and Toxicity Based on Chemical Structure Information
We develop quantitative structure–activity relationship (QSAR) models to predict the biological activities and toxicities of pharmaceuticals and chemicals in silico from their chemical structures. By leveraging machine learning and deep learning techniques with molecular descriptors, fingerprints, and our in-house substructure descriptors, we aim to build predictive models that are both highly accurate and interpretable.
-
Pharmacovigilance Research Using Adverse Event Databases
We detect adverse drug reaction signals by applying disproportionality analyses and statistical models to spontaneous reporting databases such as the FDA's FAERS and the PMDA's JADER. By elucidating the occurrence patterns of clinically important adverse reactions—including severe cutaneous adverse reactions, drug-induced liver injury, and adverse events associated with cancer chemotherapy—we contribute to improving the safety of pharmaceuticals.
-
Analysis of Toxicity Mechanisms Based on Adverse Outcome Pathways (AOPs)
We analyze toxicity mechanisms by integrating AI and QSAR technologies within the Adverse Outcome Pathway (AOP) framework, which describes the sequence of biological processes from molecular initiating events (MIEs) to adverse outcomes. By linking epidemiological signals of adverse drug reactions with their molecular-level mechanisms, we aim to realize mechanism-based toxicity prediction and safety evaluation.