Computational Metabolomics: From Cheminformatics to Machine Learning (Dagstuhl Seminar 20051)

Abstract

Dagstuhl Seminar 20051 on Computational Metabolomics is the third edition of seminars on this topic and focused on Cheminformatics and Machine Learning. With the advent of higher precision instrumentation, application of metabolomics to a wider variety of small molecules, and ever increasing amounts of raw and processed data available, developments in cheminformatics and machine learning are sorely needed to facilitate interoperability and leverage further insights from these data. Following on from Seminars 17491 and 15492, this edition convened both experimental and computational experts, many of whom had attended the previous sessions and brought much-valued perspective to the week’s proceedings and discussions. Throughout the week, participants first debated on what topics to discuss in detail, before dispersing into smaller, focused working groups for more in-depth discussions. This dynamic format was found to be most productive and ensured active engagement amongst the participants. The abstracts in this report reflect these working group discussions, in addition to summarising several informal evening sessions. Action points to follow-up on after the seminar were also discussed, including future workshops and possibly another Dagstuhl seminar in late 2021 or 2022.