New molecules for new pharmaceuticals using artificial intelligence

A major international scientific collaboration has for the first time validated the effectiveness of a tool that uses artificial intelligence to design new molecules with potential pharmacological use. The study - published in Scientific Reports - involved more than 250 research institutions in 30 countries, including the University of Bologna with the departments of Medical and Surgical Sciences and Pharmacy and Biotechnology.

The new technology could be a much more effective alternative to High-throughput Screening (HTS), the most commonly used process today for analysing large numbers of molecules and identifying potential new drug functions among them.

The AtomNet technology platform, developed by Atomwise, is the first neural network designed to predict the bioactivity of small molecules in the drug discovery process. Thanks to its virtual approach to high-throughput screening, the system is able to search a library of many billions of synthesised compounds to find new solutions in a vast unexplored chemical space. The researchers used the new artificial intelligence-based system to carry out high-throughput virtual screening - the largest and most numerous to date - analysing 318 target molecules and generating new results across all major therapeutic areas and protein classes.

"The work of developing new medicines is focused on increasingly difficult diseases, so we need new molecules with which to find new solutions, -  explains Giuseppe Gasparre, professor at the Department of Medical and Surgical Sciences and director of the Centre for Applied Biomedical Research at the University of Bologna, who is one of the authors of the study. - "An artificial intelligence-based approach like the one we have validated in this study could open up numerous new possibilities and play a key role in the development of new medicines”.

The study was published in Scientific Reports under the title 'AI is a viable alternative to high throughput screening: a 318-target study.' Participating for the University of Bologna were Giuseppe Gasparre and Monica De Luise from the Department of Medical and Surgical Sciences, together with Anna Maria Porcelli and Luisa Iommarini from the Department of Pharmacy and Biotechnology.