Bishnu Sarker, Ph.D., assistant professor, computer science and data science, has received a National Science Foundation (NSF) Historically Black Colleges and Universities – Excellence in Research (HBCU-EiR) grant. The $533,752 award will fund the project, “Developing a Knowledge Graph Driven Integrative Framework for Explainable Protein Function Prediction via Generative Deep Learning.”
“I am so grateful to the NSF for supporting my research applying artificial intelligence to predict protein functions,” says Dr. Sarker. “The tool I plan to develop in this project can help further the work in drug discovery, particularly with repurposing existing drugs for new treatments.”
Dr. Sarker’s research seeks to improve on previous work to predict protein function, which can lead to important insights about disease prevention, diagnosis and treatment.
His goal is to create a new software tool that will serve as a protein function prediction pipeline.
“I plan to use generative deep learning approach, based on a knowledge graph connecting heterogeneous biological concepts using multitudes of relations pertaining to proteins,” says Dr. Sarker. “A large language model will be applied to the multi-omics data to use that information to gain a better understanding of the proteins and their functions.”
In addition to creating a robust, accurate tool, Dr. Sarker seeks to avoid black box concerns often associated with artificial intelligence. He will work to make the software explainable so that the steps in its decision making will be clear. It will also be reproducible so that others can use it and achieve the same results.
Aize Cao, Ph.D., associate professor, biomedical data science, is co-principal investigator. The NSF grant will also fund multiple graduate assistants, providing an important opportunity for minority students at Meharry to pursue AI/ML, quantitative genomics and data science research.