Building a system to support a data-driven drug discovery DMTA cycle
Akimi HORI *, Seisuke TAKIMOTO, Keisho OKURA, Satoki DOI, Teruhiko INOUE
Central Pharmaceutical Research Institute, JAPAN TOBACCO INC.
[Background and Purpose]
A critical component of research is deciding how to manage, accumulate, and utilize data —including both actual measured data and virtual data, e.g., predicted values and ideas. Both data types are crucial, and contain abundant information related to project advancement and organizational direction setting.
[Method]
In Japan Tobacco Inc., challenges have arisen in utilizing actual measurement data and prediction technologies based on it. Our team has conducted pre-synthesis predictions of various compound properties, to utilize actual measurement data for more productive drug discovery process. These improvements in the drug discovery process have proven to be useful. However, they have prompted a need to construct a comprehensive system of interfaces and databases, to facilitate greater routine implementation of predictive calculations and use of predicted values within projects, and to accumulate data obtained from project operations for retrospective analysis.
[Result and Discussion]
In this presentation, we will introduce the system built to challenge Japan Tobacco's predict-first design-make-test-analyze (DMTA) cycle, along with the underlying concepts, including technical details.