It can take billions of dollars and decades to develop life-saving drugs, but researchers at the University of Central Florida want to accelerate that process with a new artificial intelligence-based drug screening technique they’ve developed.
Using a method that models drug-target protein interactions using natural language processing techniques, researchers achieved up to 97% accuracy in identifying promising drug candidates. The results were recently published in the journal Briefings in bioinformatics.
The technique represents drug-protein interactions by words for each protein binding site and uses deep learning to extract the features that drive the complex interactions between the two.
“With the increasing availability of AI, this has become something that AI can handle,” says study co-author Ozlem Garibay, an assistant professor in UCF’s Department of Industrial Engineering and Management Systems. “You can try so many variations of proteins and drug interactions and see which ones are more likely to bind or not.”
The model they developed, called AttentionSiteDTI, is the first to be interpretable in the language of protein binding sites.
The work is important because it will help drug developers identify critical protein binding sites along with their functional properties, which is key to determining a drug’s efficacy.
Researchers achieved the feat by developing a self-awareness mechanism that lets the model learn which parts of the protein interact with the drug compounds, while achieving state-of-the-art predictive power.
The mechanism’s self-awareness ability works by selectively focusing on the most relevant parts of the protein.
The researchers validated their model using laboratory experiments that measured binding interactions between compounds and proteins, and then compared the results to those that their model had predicted computationally. As drugs to treat COVID remain of interest, experiments also included testing and validating drug compounds that would bind to a spike protein of the SARS-CoV2 virus.
According to Garibay, the high agreement between the laboratory results and the computational predictions highlights the potential of AttentionSiteDTI to pre-screen potentially effective drug compounds and accelerate the discovery of new drugs and the repurposing of existing drugs.
“This high-impact research was only possible because of the interdisciplinary collaboration between materials engineering and AI/ML and computer scientists to address discoveries related to COVID,” says Sudipta Seal, co-author of the study and chair of UCF’s Department of Materials Science and Engineering.
Mehdi Yazdani- Jahromi, a PhD student at UCF’s College of Engineering and Computer Science and lead author of the study, says the work heralds a new direction in drug screening.
“This allows researchers to use AI to more accurately identify drugs to respond quickly to emerging diseases,” says Yazdani- Jahromi. “This method also allows researchers to identify the best binding site of a viral protein, which they can focus on in drug design.”
“The next step in our research will be to develop novel drugs that harness the power of AI,” he says. “Of course, this can be the next step in being prepared for a pandemic.”
The research was funded by UCF’s internal AI and big data seed funding program.
Also among the study’s co-authors were Niloofar Yousefi, a postdoctoral fellow in UCF’s Complex Adaptive Systems Laboratory at UCF’s College of Engineering and Computer Science; Aida Tayebi, PhD student in UCF’s Department of Industrial Engineering and Management Systems; Elayaraja Kolanthai, postdoctoral researcher at UCF’s Department of Materials Science and Engineering; and Craig Neal, a postdoctoral researcher in UCF’s Department of Materials Science and Engineering.
Garibay holds a PhD in Computer Science from UCF and joined UCF’s Department of Industrial Engineering and Management Systems in 2020, which is part of the College of Engineering and Computer Science. Previously, she worked in information technology for UCF’s Office of Research for 16 years.
Materials provided by University of Central Florida. Originally written by Robert Wells. Note: Content can be edited for style and length.