They find a “shortcut” to increase drug discovery capabilities.

Most drugs are small molecules that affect the activity of human proteins. And ligands are a chemical union between both things. However, for most proteins we still do not know which ligands are effective. In particular, it is surprising that for For more than 80% of all proteins, no binder has yet been discovered. molecules small.

This is important because very often ligands represent the starting point for drug development. So, we can say that this knowledge gap in many cases directly leads to our inability to create new drugs. Something that at the same time hampers the development of therapeutic strategies and prevents the acquisition of new biological knowledge about health and disease.

Now researchers from the CeMM Research Center for Molecular Medicine of the Austrian Academy of Sciences, in collaboration with the pharmaceutical company Pfizer, have created a new method measure the binding activity of hundreds of small molecules to thousands of human proteins. This new large-scale study has created a “rich catalog of tens of thousands of ligand-protein interactions” that can now be explored to develop chemical and therapeutic tools and will provide a “shortcut” to drug discovery. .

A new model based on machine learning and artificial intelligence allows for impartial forecasting how small molecules interact with all the proteins present in living human cells. The results of the study were published this Thursday in the journal The scienceand, as the authors themselves explain, all obtained data and models are freely available to the scientific community.

In their study, the team led by CeMM principal investigator Georg Winter demonstrated this by developing small molecules that bind cellular transporters.components of the cellular degradation mechanism and insufficiently studied proteins involved in cellular signal transduction.

“We were surprised to see how artificial intelligence and machine learning could improve our understanding of the behavior of small molecules in human cells. We hope that our catalog of interactions between small molecules and proteins and associated artificial intelligence models can now provide the fastest path to drug discovery,” Winter said.

To maximize potential impact and usefulness to the scientific community, all data and models are available free of charge through a web application. “This has been an exceptional partnership between industry and academia. We are pleased to present the results which were obtained through three years of close cooperation and teamwork between groups. “This was a great project,” explained Dr. Patrick Verhost, vice president and head of drug development at Pfizer.