The protein design assistant from the Leipzig-based start-up AI-Driven Therapeutics (AI-DT) is an AI-based platform that is designed to work autonomously. Tasks that would normally take weeks could be solved in just a few hours. The programme draws on large amounts of data and scientifically validated models. “Simply put, we are developing a kind of ChatGPT for protein design: a platform that creates an autonomous co-scientist by providing specific software tools – a collegial partner that is more than just a tool,” explains pharmacist Liessmann. Their mission is to make their AI-supported protein design methods accessible to small and medium-sized companies that do not have state-of-the-art AI technologies and to significantly accelerate their research and development processes.

In order to optimise the binding of an antibody, for example, various amino acids in this protein can be exchanged. There are 20 natural amino acids available per mutation, which makes this task very complex from a combinatorial point of view. With six positions to be mutated, there are already 64 million combinations. If the researchers were to test all the mutants in the laboratory, this would be extremely time-consuming. “Our co-scientist can predict what is most logical and fits together within a few hours on the PC,” explains Liessmann.

Some of the methods for this were developed by the Nobel Prize winner in chemistry David Baker. This biochemist, in turn, was the Humboldt Professor Jens Meiler from Leipzig who completed his postdoc 20 years ago. The current head of the IWE supports the young researchers as a professional mentor and networks them with relevant stakeholders. The connection to Leipzig University and Jens Meiler's working group should remain close. Among other things, SECAI offers the start-up a valuable basis for interdisciplinary exchange between AI method development and biotechnological/biomedical research.

For the project, the team has been funded with the EXIST start-up grant from the Federal Ministry for Economic Affairs and Climate Protection totalling 120,000 euros since October. Within this year, the founding team now wants to find pilot customers on which their AI scientist can practise and improve his processes and have developed the co-scientist to such an extent that it can be used on the market by autumn 2025.

[Excerpts from the Universität Leipzig press release; Auf der Überholspur zu neuen Arzneimitteln]