Florida scientists train AI to identify drugs' impact on cellular targets

In Jupiter, Florida, a team of researchers led by neuroscientist Kirill Martemyanov, Ph.D. from The Herbert Wertheim UF Scripps Institute for Biomedical Innovation & Technology has successfully trained an AI system to predict how drugs will impact the largest family of cellular targets with over 80% accuracy. This cutting-edge research has the potential to revolutionize the field of precision medicine.

Traditionally, prescribing medication has been a one-size-fits-all approach, with doctors relying on trial and error to determine which drugs will work for individual patients. However, this approach can lead to ineffective or harmful outcomes as people have significant genetic variability in their cell receptors. To address this, Martemyanov's team utilized molecular tracking technology and AI to profile the action of over 100 cellular drug targets, including genetic variations.

The researchers gathered data from a decade of experimentation and an extensive collection of information on G protein-coupled receptors (GPCRs) behavior, which are responsible for a third of all drug responses. GPCRs play a vital role in pain relief, allergies, blood pressure regulation, and other biological activities. By training the AI algorithm using this comprehensive dataset, the scientists achieved an impressive 80% accuracy in predicting how GPCRs would respond to drug-like molecules.

Martemyanov emphasized the importance of understanding the complexity of GPCRs, stating, "We all think of ourselves as more or less normal, but we are not. We have tremendous variability in our cell receptors. If doctors don’t know what exact genetic alteration you have, you just have this one-size-fits-all approach to prescribing, so you have to experiment to find what works for you."

The team's research also led to the discovery of surprising differences in how mutated GPCRs responded to stimuli. This additional knowledge has opened up new possibilities for tailored prescriptions and the design of truly personalized medications.

Martemyanov credited the collaboration with computational protein designer Bruno E. Correia, Ph.D., and researcher Ikuo Masuho, Ph.D., as instrumental in the development of the AI algorithm. Their combined expertise and a decade-long dataset helped the researchers overcome the previous lack of accurate and detailed GPCR activity information.

The successful outcomes of this study could have significant implications for drug development and patient safety. By adopting a more sophisticated understanding of GPCRs and their interactions with drugs, pharmaceutical companies could create safer medications more quickly and at a lower cost. The next step for the research team is to investigate how individual genetic variations affect the response to GPCR-acting drug-like compounds.

"Our ultimate goal is to predict how individual genetic variations respond to drugs, allowing for the custom tailoring of prescriptions and paving the way for precision medicine," said Martemyanov.

The study, titled "Rules and mechanisms governing G protein coupling selectivity of GPCRs," was authored by Ee Von Moo, Xiaona Li, and Hideko Wakasugi-Masuho from The Wertheim UF Scripps Institute, Ryoji Kise and Ryosuke Tany from Sanford Research, and Pablo Gainza from the École Polytechnique Fédérale de Lausanne and Swiss Institute of Bioinformatics in Lausanne, Switzerland.

The research received funding from the National Institutes of Health through grants DA036596 and MH105482, as well as from the Swiss National Science Foundation and startup funding from Sanford Research.

Precision medicine is making significant progress, which has the potential to greatly impact patient care and improve health outcomes. This innovative AI-powered approach creates new possibilities for personalized drug treatment, bringing us closer to a future where medications are customized based on an individual's genetic makeup.