Traditional drug discovery and its safety and functional evaluation relies heavily on the concept of trial and error, making the process slow, expensive and limited in its adaptability. Modern computational methods have shown the potential to generate problem specific compounds on demand, simulate the mechanism of action of a compound in a virtual space, and in addition function under the view of personalized medicine.

However, the application of these methods outside of a research setting is often challenging due to their lack in explainability, traceability, and need to function in a heavily regulated space, such as the pharmaceutical space. Given the growing availability of ML-driven methods in the drug discovery and evaluation space this session aims at bringing together experts from different sub-domains of the research and application space to:

1. Discuss the current possibilities of ML based methodologies in drug discovery and their evaluation
2. Point out the unique challenges these approaches face under EU/ Danish law, generate constructive discussion on current challenges and discuss potential research directions in order to get these promising methods from research and into application domain.

Program

The first part of the session consists of invited talks (45 minutes in total), functioning as a starting point and introduction for the panel discussion.

The second part of the session is an open panel discussion (45 minutes). The panel discussion is an opportunity for the audience to engage with the experts but also an opportunity for the audience to pitch in with their point of view as well as to challenge the panelists and audience in discussing the current gap between possibilities in AI methodologies and their practical application in a regulated field.

AI transforming therapeutic discovery into therapeutic design
Timothy Jenkins, Assistant Professor (speaker, panelist), Department of Bioengineering, DTU

Considering application domain when using AI/ML
TBA

The Role of AI in Drug Discovery: 5 Key Lessons from Industry Experience
Paolo Marcatili, Head of Antibody Design, prev. Assoc. Prof. DTU Health (speaker, panelist) Head of Antibody Design, Novo Nordisk

Quality and compliance of ML/AI in the pharmaceutical industry
Naram El-Shamary, Co-founder and CEO, (panelist), ImproveMatic

Organizers

• Line H. Clemmensen, Professor (lkhc@dtu.dk), KU & DTU Compute (STAT), Lead Organizer
• Alisa Pavel, Postdoc (alpav@dtu.dk), DTU Compute (STAT)
• Manja Grønberg, Postdoc (mgegr@dtu.dk), DTU Compute (STAT)
• Mathies B. Sørensen, Postdoc (mabso@dtu.dk), DTU Compute (STAT)
• Mikkel N. Schmidt, Associate Professor (mnsc@dtu.dk), DTU Compute (COGSYS)

Level

Introductory