What are the unique aspects or strengths of your research?

Modeling and prediction using machine learning and artificial intelligence (AI) have been highly successful when high-quality, large-scale data, such as protein 3D structures, are available. However, in many areas of biology, including drug discovery, data can often be scattered or lack sufficient quality, making it challenging to fully leverage AI. Our research focuses on building robust AI systems based on data by creating databases with our own data integration tools and developing efficient curation systems. We believe our strength comes from this data-centric approach and our ability to create predictive models for a wide range of topics including nanoparticles, nucleic acid drugs, and drug metabolism and pharmacokinetics.

How do you think the results of your research will benefit society or industry?

One major reason why developing new drugs takes a significant amount of time and money is that candidates identified in cell and animal experiments do not always result in effective treatments for humans. Since direct experimentation on humans is not feasible, it is crucial to use indirect information to enhance our understanding of biological processes. Integrating and modeling various types of biological data plays a key role in this endeavor. We believe that our databases and predictive models, for drug metabolism and pharmacokinetics and for gut microbiome and lifestyle-related data, can help not only in controlling diseases through medications but also in promoting a healthier society by supporting personalized health maintenance.

How is data science utilized in your research?

Almost every part of our research makes use of data science techniques. This includes curating data—collecting and integrating the right data for each problem to create high-quality training datasets. We then visualize and analyze patterns in the data, building predictive models using various machine learning methods such as deep learning and developing software to share our databases and tools as web applications.

Please share examples of collaborative research or the potential for future collaboration.

By collaborating with pharmaceutical and technology companies, we are developing new methods to improve the efficiency of protein design and drug development. In the past, we established a data-sharing framework by bringing together a consortium of pharmaceutical companies. Additionally, we implemented our research into practical applications by licensing our software and databases to IT companies for their commercialization. We aim to continue a wide range of collaborations, including academic research on understanding biological processes and disease mechanisms, as well as industry partnerships focused on the discovery of new drug targets.

What are the prospects and goals for your research?

We aim to provide technical solutions to address questions such as how to collect and share computer-friendly data for AI applications, as well as how to design incentives that motivate data producers to willingly share their data. There are still many challenges in connecting molecular-level analysis, like proteins, with higher-level biological phenomena like cells and whole organisms. By integrating diverse biological data, we tackle this issue—for example, by linking molecular modeling of drug-metabolizing enzymes to individual differences in drug response and gaining deeper insights into these complex biological processes.

Research papers

1. Wang et al., Biological age prediction using a DNN model based on pathways of steroidogenesis, Science Advances, in press (2024).
   https://doi.org/10.1126/sciadv.adt2624
2. K. Koyama, K. Hashimoto, C. Nagao and K. Mizuguchi, Attention network for predicting T-cell receptor–peptide binding can associate attention with interpretable protein structural properties, Front. Bioinform., 3, 1274599 (2023).
   https://doi.org/10.3389/fbinf.2023.1274599
3. H. Kawashima, R. Watanabe, T. Esaki, M. Kuroda, C. Nagao, Y. Natsume-Kitatani, R. Ohashi, H. Komura and K. Mizuguchi, DruMAP: A novel drug metabolism and pharmacokinetics analysis platform, J. Med. Chem., 66(14), 9697-9709 (2023).
   https://doi.org/10.1021/acs.jmedchem.3c00481
4. R. Watanabe, K. Hashimoto, K. Higashisaka, Y. Haga, Y. Tsutsumi and K. Mizuguchi, Evidence-based prediction of cellular toxicity for amorphous silica nanoparticles, ACS Nano, 17(11), 9987-9999 (2023).
   https://doi.org/10.1021/acsnano.2c11968
5. Y.-A. Chen, R. S. Allendes Osorio and K. Mizuguchi, TargetMine 2022: a new vision into drug target analysis, Bioinformatics, 38(18), 4454–4456 (2022).
   https://doi.org/10.1093/bioinformatics/btac507