AI-based screening method could speed up new drug discovery
Developing life-saving drugs can take billions of dollars and decades of time, but University of Central Florida researchers aim to speed up that process with a new artificial intelligence-based drug screening process they’ve developed .
Using a method that models interactions between drugs and target proteins using natural language processing techniques, researchers have achieved up to 97% accuracy in identifying promising drug candidates. The results were recently published in the journal Bioinformatics Briefings.
The technique represents drug-protein interactions through words for each protein binding site and uses deep learning to extract the features that govern the complex interactions between the two.
“With AI becoming more available, it’s become something AI can tackle,” says study co-author Ozlem Garibay, an assistant professor at UCF. Department of Industrial Engineering and Management Systems. “You can try many variations of proteins and drug interactions and find out which ones are most likely to bind and which ones aren’t.”
The model they developed, known as AttentionSiteDTI, is the first to be interpretable using the language of protein binding sites.
The work is important because it will help drug designers identify critical protein-binding sites and their functional properties, which is critical in determining whether a drug will be effective.
The researchers achieved this feat by designing a self-attention mechanism that allows the model to learn which parts of the protein interact with drug compounds, while achieving peak prediction performance.
The mechanism’s self-attention ability works by selectively focusing on the most relevant parts of the protein.
The researchers validated their model using lab experiments that measured the binding interactions between compounds and proteins, then compared the results with those their model predicted computationally. As drugs to treat COVID are still of interest, the experiments also included the testing and validation of drug compounds that would bind to a spike protein of the SARS-CoV2 virus.
Garibay says the high agreement between lab results and computational predictions illustrates the potential of AttentionSiteDTI to prescreen potentially effective drug compounds and accelerate the exploration of new drugs and the repurposing of existing ones.
“This high-impact research was only possible through an interdisciplinary collaboration between materials engineering and AI/ML and computer scientists to address COVID-related discovery,” says Sudipta Seal, co-author of the study and president of UCF. Department of Materials Science and Engineering.
Mehdi Yazdani-Jahromi, PhD student at UCF College of Engineering and Computer Science and the study’s lead author, says the work introduces a new direction in drug prequalification.
“This allows researchers to use AI to identify drugs more accurately to respond quickly to new diseases,” Yazdani-Jahromi says. “This method also allows researchers to identify the best binding site of a viral protein to focus on in drug design.”
“The next step in our research will be to design new drugs using the power of AI,” he says. “That can naturally be the next step in preparing for a pandemic.”
The research was funded by UCF’s internal AI and Big Data funding program.
Study co-authors also included Niloofar Yousefi, a postdoctoral research associate at UCF Complex Adaptive Systems Laboratory at UCF College of Engineering and Computer Science; Aida Tayebi, PhD student in the Department of Industrial Engineering and Management Systems at UCF; Elayaraja Kolanthai, postdoctoral research associate in the Department of Materials Science and Engineering at UCF; and Craig Neal, postdoctoral fellow in UCF’s Department of Materials Science and Engineering.
Garibay received her doctorate in computer science from UCF and joined the Department of Industrial Engineering and Management Systems at UCF, part of the College of Engineering and Computer Science, in 2020. Previously, she worked for 16 years in information technology for UCF. Research Office.
Article Title: AttentionSiteDTI: An Interpretable Graphical Model for Drug-Target Interaction Prediction Using NLP Sentence-Level Relationship Classification