Abstract: RAS proteins (KRAS, NRAS, and HRAS) are the most extensively studied set of mutationally activated oncogenes. Yet we do not completely understand the structural impact of RAS mutations in cancer. As of August 2020, there are 333 experimentally solved wild-type (WT) and mutated RAS structures in the Protein Data Bank (PDB), comprising 175 HRAS, 155 KRAS, and 3 NRAS structures. This growing structural ensemble provides a valuable resource to discover novel insights into RAS activity through statistical analyses that enable quantitative determination of biological correlates. In recent years, NMR studies and molecular dynamic (MD) simulations (using experimental structures as templates) have shown that the conserved RAS switch 1 (SW1) and 2 (SW2) regions display dynamic conformational behaviors that modify RAS activity. Conformational changes in SW1/SW2 facilitate the concerted binding of regulator and effector proteins at these regional interfaces, in turn promoting proper switching of RAS proteins from an active GTP-conformer to an inactive GDP-conformer. In addition, biochemical studies have provided evidence for the existence of a GTP-bound RAS homodimer required for activation of certain signaling pathways. While we know that RAS homodimerization occurs in a GTP-bound state, we do not know if a specific RAS conformational arrangement is required for RAS homodimer formation and associated protein-protein interaction (PPI) events. Further complicating this issue, there is no consolidated understanding of how mutations on RAS proteins may shift SW1/SW2 conformation in ways favoring or disfavoring RAS PPIs, including RAS homodimerization. The objective of this proposal is to create a unified RAS structural classification to assess the impact of oncogenic mutations on KRAS, HRAS, and NRAS, ligands (GTP, GDP, inhibitors, or none), and PPIs/homodimerization. In preliminary work, I created a unified RAS structural nomenclature based on clustering SW1/SW2 conformations across experimental structures of KRAS, HRAS, and NRAS. I have also performed a pan-cancer analysis of 18,841 RAS missense mutations from a cohort of 100,707 patients, providing the most comprehensive existing resource for KRAS, NRAS, and HRAS mutational patterns in human tumors. In Aim 1, I will compare the conformational and PPI preferences of RAS mutated and WT forms with experimentally solved structures. Following this, in Aim 2, I will predict the conformational and PPI preferences of RAS mutated and WT forms by examining energy distributions of generated structural ensembles. If we can reproduce the effects of some known mutations, then we can confidently predict the consequence for novel mutations that have not been experimentally studied. In completing this proposal, I will present the unified RAS structural nomenclature, including the determined impact of RAS mutations, in a database that will be continually updated upon solving of new experimental structures. This work...