P03-21
Data utilization and DX talent development on in-house KNIME platform
Toshiyuki OHFUSA *1, Takanobu ARAKI2, Ikumi KURIWAKI2, Ayato SUGIYAMA1, Kazuya NAGAOKA1, Kenichi MORI1, Takamune YAMAMOTO3, Kenji NEGORO2, Kota TOSHIMOTO4, Shinji SOGA5, Takuya SHIMOMURA6, Tomomi YOKOYAMA7, Hiroko TAMURA6
1Modality Informatics, ResearchX, DigitalX, Astellas Pharma Inc.
2Platform Sciences & Modalities, Discovery Intelligence, Applied Research & Operations, Astellas Pharma Inc.
3Research Informatics, ResearchX, DigitalX, Astellas Pharma Inc.
4Systems Pharmacology, Advanced Translational Science & Management, Non-Clinical Biomedical Science, Applied Research & Operations, Astellas Pharma Inc.
5Biologics Engineering, Discovery Intelligence, Applied Research & Operations, Astellas Pharma Inc.
6Pharmaceutical Developability Labs, CMC Research, Astellas Pharma Inc.
7Technology Exploration 1, CMC Research, Astellas Pharma Inc.
( * E-mail: toshiyuki.ohfusa@astellas.com )
In recent years, various modalities have been developed for drug discovery approaches which target undruggable targets. These modalities include "bRO5" (beyond rule of 5) chemical modalities, such as bifunctional compounds as protein degraders and peptide molecules, as well as various biological modalities. Additionally, coupled with advances in automation equipment and the use of new technologies in the drug discovery field, such as generative AI, the amount of data generated at each step of the drug discovery process is steadily increasing.
The increased data can be utilized for molecular predictions and simulations, leading to reduced operating costs. Moreover, enabling data-driven decision-making to improve the quality and efficiency of drug discovery is believed to expedite the delivery of new drugs to patients.
However, many wet lab researchers are not accustomed to handling large amounts of data, especially in CUI (Character User Interface) environments. Data interface issues pose a hurdle to learning in various fields and contribute to inefficiencies in data utilization within each laboratory. Therefore, our cross-departmental DX promotion team is building a foundation platform for DX activities to increase the number of individuals capable of handling data in various wet lab tasks. This platform utilizes KNIME, a GUI (Graphical User Interface)-based no-code/low-code tool that is easy for beginners to understand and learn. The developed tool can be shared with other members as a web application by deployment on the KNIME Server platform installed on a shared server.
Furthermore, as an educational activity for members of each department, the DX team has established an in-house community to facilitate study sessions, information sharing, and provide a platform for questions and consultations. KNIME enables researchers in each department to utilize data effectively by leveraging their domain knowledge. Consequently, they can approach problem-solving from a field perspective in the development of various modalities, leading to significant improvements in decision-making speed and work efficiency. As part of data utilization efforts, automation of regular report creation within each organization, automatic data aggregation across organizations, and the development of prediction models have increased, in turn resulting in significant time and effort savings. Additionally, this platform is widely used not only in the discovery research department but also in the CMC research department and manufacturing department, with an increasing number of application developers in each department.
In this poster presentation, we will showcase the utilization of data and DX human resource development using the in-house KNIME platform.
