Computer scientist Lydia Kavraki of Rice University’s Brown School of Engineering has won a prestigious National Institutes of Health U01 grant to develop a new approach to model and analyze protein-ligand interactions in cancer research.
The end goal is to create a proteomics toolkit, PROTEAN-CR, focused on the structural analysis of protein-ligand interactions. Researchers will use PROTEAN-CR to understand key biological mechanisms of cancer as well as to suggest novel cancer therapies. Pilot projects will include peptide-based cancer vaccination and analysis of mutations in the context of T-cell based immunotherapy.
The three-year, $1.2 million grant from the National Cancer Institute (NCI) will advance Kavraki's ongoing collaboration with co-investigator Gregory Lizee at the University of Texas MD Anderson Cancer Center.
In cancer, proteins can suffer modifications that favor the maintenance and proliferation of malignant cells. One way to fight cancer is using ligands with anti-tumor properties to inhibit these proteins. But the discovery of molecules with anticancer properties isn't easy. There are hundreds of thousands of different proteins and possible ligands to assess. Proteins and ligands can assume different three-dimensional conformations, and even the same protein can have multiple mutations; both these issues affect protein-ligand binding. Kavraki said PROTEAN-CR is needed because of these challenges and a persistent knowledge gap with structural analyses of protein–ligand interactions.
Kavraki, Rice's Noah Harding Professor of Computer Science, a professor of bioengineering, mechanical engineering and electrical and computer engineering, and director of the Ken Kennedy Institute, said the long-term goal of the U01 grant is to enable broad structural analysis of protein-ligand interactions so cancer researchers can mix, match and test small anti-tumor molecules for personalized cancer therapies.
PROTEAN-CR will also allow researchers to manipulate the 3D structures of known molecules and their possible forms, making it easier for researchers to screen possible protein-ligand complexes and predict how they will bind to and destroy tumor cells. To assess the best binding modes, machine learning methods will be used to create new scoring functions. PROTEAN-CR also will be linked to publicly available biological databases to retrieve updated information on protein mutations and modifications.
Kavraki said her unified data science-inspired approach will accelerate cancer research by complementing wet-lab and clinical studies. Additional collaborators include Dinler Antunes at the University of Houston and Jin Wang at Baylor College of Medicine.
"There is a real gap in incorporating large-scale structural analysis to understand the role of proteins and protein-ligand interactions in complex diseases such as cancer," Kavraki said. "Our work will fill this gap and complement the tools that are currently in development through NCI's informatics program."
Kavraki's lab has already developed some PROTEAN-CR core functions through a prototype web server being tested by MD Anderson researchers for drug discovery and immunotherapy applications. Since March 2017, it has had more than 9,752 unique users from 109 countries.