Study shows that diverse artificial intelligence models are successful in identifying approved drugs to be repurposed against drug-resistant bacterial pathogens.
Streptococcus pneumoniae is a bacterial pathogen that causes life-threatening infections, including pneumonia and meningitis. These infections are commonly treated with antibiotics, but antibiotics are becoming less effective against S. pneumoniae over time due to changes in the pathogen that cause drug resistance. This phenomenon, called antimicrobial resistance, is resulting in significant healthcare costs and deaths around the world. In a new study, published by Wiley in Advanced Science, researchers used artificial intelligence (AI) to identify approved drugs that may be effective against S. pneumoniae.
Drug repurposing, or the use of drugs with known safety profiles for different purposes, saves time and cost while reducing risk. This process is even more efficient with the use of computational strategies to select promising candidates. AI has been used to predict drugs for repurposing against other pathogens on the World Health Organization (WHO)'s antibiotic development priority list. Now, researchers have used enhanced learning approaches and diverse models to screen a drug library for S. pneumoniae, another of the WHO's priorities for antimicrobial resistance.
Using a dataset of molecules shown to be active against S. pneumoniae and a larger dataset of molecules inactive against another drug-resistant bacteria, the scientists trained three different learning algorithms: ensembles of decision trees, ensembles of graph neural networks, and ensembles of sequence-based transformers (each previously pre-trained with hundreds of millions of molecules). These different AI models were used to test almost 7,000 candidate drugs for inhibition of S. pneumoniae.
Of these nearly 7,000 candidates, 11 were selected for experimental validation. Nine of the compounds were able to inhibit the growth of S. pneumoniae. One of the two most potent repurposed drugs was effective even against drug-resistant strains of S. pneumoniae. The authors discussed that using three diverse AI models complemented each other in the selection of candidate drugs, making them more efficacious together than individually.
These findings underscore the benefits of using AI models to identify novel drugs and rationally select drugs with the highest potential for repurposing. In a field like antimicrobial resistance, where there is an urgent need for effective compounds, AI-driven computational drug repurposing is a streamlined and affordable strategy with demonstrated success.
"AI-guided drug repurposing has become a powerful strategy for combating antimicrobial resistance, particularly for pathogens for which some active molecules are already known and can be leveraged as a training or fine-tuning dataset," said senior author Pedro J. Ballester, Associate Professor at the Imperial College London and Wolfson Fellow of the Royal Society in the United Kingdom.
Additional information
