Virus-Inspired Declarative Geometric Computation

University Of Florida, Gainesville FL

Investigators

Abstract

EIA-0218435 Meera Sitharam University of Florida Virus-Inspired Declarative Geometric Computation The goal of the project is to develop geometric computational models and tools for virus assembly from their constituent proteins, and virus crystal packing. Furthermore, inspiration of the above processes is being used to rethink computationally tractable declarative geometry (DG), defined as the intuitive, constraint-based representation and efficient realization of composites of simple interacting geometric objects, starting from a declarative specification of the composite's properties. In particular, a new game-theoretic constraint model is being developed for the underlying class of algebraic-geometric computations and corresponding algebraic varieties. Existing software in the form of the PI's geometric constraint solver FRONTIER is forming the base for implementing the new computational framework. The new virus computational models is used for the studying the following unanswered questions on carefully chosen, geometrically significant viruses: (a) the isolation of crucial geometric events during assembly (helpful for disrupting assembly); (b) the isolation of assembly events - such as molecular conformational changes - that require the involvement of viral genomic material, (helpful for understanding DNA-protein interactions); and (c) the isolation of key geometric events during virus crystallization (as an idealized version of molecular crystallization). The new DG virus models is being refined and validated by checking consistency with known behavior of viruses and their constituent proteins during assembly and crystallization. A small number of other highly focused experiments; selective X-ray crystallography and/or cryoelectron microscopy is being performed. As a significant player to help with both of the above goals, use the distinctive Maize-streak virus (MSV) will be used, whose structure and properties are particularly suited to goals of the project. A comparison of the new geometric virus models with other geometry-based computational virus models is being made.