PROJECT SUMMARY/ABSTRACT The central objectives of our MIRA research program are "development" and "validation" of methodologies to algorithmically encode underlying physical observables to improve design of small organic molecules for a biological target. Computational modeling at the atomic-level empowers understanding of the factors that drive molecular recognition, and "application" to real world systems enables testable predictions. Grounded in strong results, data, and demonstrated productivity, we hypothesize that major gaps in the field (e.g. design of ligands with optimal target complementarity, sampling/searching challenges, protein specificity, incorporating cheminformatics during design, lack of intelligent automated-methods for refinement, lack of methods for covalent de novo design, difficult to use software) can be bridged through forward-thinking design of new computational tools. As the lead developer of the widely-used program DOCK6, in major undertakings, we are developing powerful paradigms for "de novo design" and "molecular evolution" that logically drive ligand construction to drug-like space using clever scoring functions and cheminformatics descriptors simultaneously. Our platforms remove the limitation of only considering ligands that are preconceived, and promote the design of compounds that are highly optimized and "specifically tailored" to the protein(s) of interest. Our approach is centered around fragment-based assembly and searching using customizable building block libraries for "from- scratch" construction and "refinement". Our advances are made available to the research community through regular public release (we led the last seven DOCK6 releases 6.4-6.10) along with validation databases and user-friendly tutorials. A new webserver will enable novice, intermediate, and expert users alike more direct access to our tools. Our expected outcomes are cutting-edge software and identification of highly specific and optimized ligands. The proposal is framed around 4 fundamental questions: (Q1) Can the principles that drive molecular recognition be captured at the atomic level and used to design improved software for more accurate prediction of geometry and energy? (Q2) Can ligand growth be propelled to highly specific regions of chemical space using clever from-scratch assembly (de novo design) and Darwinian principles (molecular evolution)? (Q3) Can more "intelligent" sampling, scoring, and searching methods be developed for identification and design of verified-active compounds in collaboration with experimentalists? (Q4) Can docking and de novo design software and protocols be designed to be more "user friendly" while not sacrificing accuracy or power? We collaborate with a network of experienced experimental labs to (i) validate theoretical predictions, (ii) answer basic research questions, (iii) provide mechanistic understanding, and (iv) identify new agents against clinically relevant targets involved in cancer ...