Future drugs could soon have fewer side effects thanks to research carried out at Coventry University and the University of Birmingham.
At Coventry University, Dr Giuseppe Deganutti has received a Research Leadership Award from the Leverhulme Trust to investigate how drug-target interactions at the molecular level can drive both therapeutic and adverse effects.
Working with collaborators at the University of Cambridge and Imperial College London, the project uses molecular dynamics simulations to study how G-protein-coupled receptors (GPCRs) interact with intracellular effectors once a drug binds to the receptor.
GPCRs are the target of most approved medicines, yet the dynamic mechanisms that link receptor activation to downstream cellular responses, including side effects, remain poorly understood.
"Small chemical modifications in drugs can lead to huge, amplified effects inside the cells," said Dr Deganutti, Assistant Professor in Coventry University’s Research Centre for Discoveries in Life Sciences.
"Being able to link structure modification on the molecule to the final effect in the cell is the sort of Holy Grail of pharmacology."
"There have been a lot of drugs on the market now for decades and many of them have adverse effects that are very strong."
"It would be nice to understand what the origins of these side effects are before trying to use the knowledge gathered to produce new drugs with improved efficacy and fewer side effects."
Unlike traditional structural biology techniques, molecular dynamics allows researchers to simulate how receptors change shape and interact with other molecules with time, providing insight into how unwanted signalling pathways may be activated.
In parallel, researchers at the University of Birmingham have developed an interpretable machine-learning framework designed to predict adverse drug reactions (ADRs) by linking real-world safety data to drug-target biology.
Published in PLOS ONE, the study, led by Joseph Roberts-Nuttall and Dr Alan M. Jones, combines high-confidence drug-target interaction data from the STITCH database with nearly three million ADR reports from the UK’s Yellow Card Scheme.
The model identifies statistically significant drug-ADR combinations and highlights the biological targets most strongly associated with specific side effects.
"One of the biggest challenges in the pharmaceutical industry is not just spotting when a side effect happens, but also understanding the reason why it occurs," said Dr Jones.
"By connecting adverse reactions back to the biological targets that each drug interacts with, this new framework brings us a step closer to safer drug development and improved patient experiences."
Using interpretable Random Forest models, the approach achieved ROC AUC scores of up to 0.94 across multiple organ-system categories, with predicted targets showing strong agreement with established disease-gene associations.
The researchers also found limited overlap between ADRs detected in clinical trials and those reported in real-world data, reinforcing the value of post-marketing datasets for safety assessment.
Together, these approaches highlight how molecular-level simulations and AI-driven safety modelling could help pharmaceutical companies identify risk earlier, refine compound design and improve confidence in drug safety across the product lifecycle.