Essential of Essential: A Variational Graph Attention Framework for Predicting Gene Essentiality and Functional Substructures

Essential of Essential: A Variational Graph Attention Framework for Predicting Gene Essentiality and Functional Substructures Understanding which genes are essential for life and why is critical for advancing biology and medicine. Essential genes, whose loss causes cell death or severe defects, are key to uncovering fundamental cellular processes and developing targeted therapies for diseases like cancer and infections. However, experimental methods to identify these genes are costly, limited by context, and unable to pinpoint the protein regions driving essentiality. Current computational models, while promising, often lack biological clarity, relying on opaque data features and failing to identify functional protein substructures. This project proposes a novel computational framework using Graph Attention Networks (GATs) to predict gene essentiality and reveal biologically meaningful protein substructures. We will develop advanced GAT models that incorporate interpretable biochemical features, such as amino acid charge and hydrophobicity, to enhance prediction accuracy and clarity. By integrating a variational graph partitioning approach, our models will identify cohesive protein modules—like catalytic sites or interaction interfaces— rather than scattered residues, aligning predictions with biological function. We will also move beyond binary essential/non-essential labels by modeling continuous fitness effects, capturing subtle gene contributions to health and disease. Our approach will use curated datasets from yeast, human, and model organisms, validated against evolutionary, disease, and functional data. Expected outcomes include more accurate, interpretable models for gene essentiality, new insights into disease mechanisms, and potential therapeutic targets. The project will engage undergraduate students in interdisciplinary research, fostering skills in biology and data science. This work will transform gene essentiality prediction, bridging computational and biological sciences to advance drug discovery, synthetic biology, and precision medicine, while training the next generation of biomedical researchers. Project Number: 1R15GM165012-01 | Fiscal Year: 2026 | NIH Institute/Center: National Institute of General Medical Sciences (NIGMS) | Principal Investigator: Jinrui Xu | Institution: HOWARD UNIVERSITY, WASHINGTON, DC | Award Amount: $566,094 | Activity Code: R15 | Study Section: Special Emphasis Panel[ZRG1 BBBT-K (86)] View on NIH RePORTER: https://reporter.nih.gov/project-details/11361339