| The research work described in this thesis covers three scales in molecular modeling: mesoscopic, microscopic, and macroscopic. In Chapter 1 (mesoscopic modeling), we generalize the discrete surface charge optimization (DiSCO) algorithm for efficient macromolecular electrostatic calculations by developing an irregular model surface building method and a DiSCO software. In Chapter 2 (microscopic modeling), we interpret an experimental finding of the effects of carcinogens on TATA/TBP bindings using molecular modeling, molecular dynamics, structural interpretations, and MM-PBSA free energy computations. In Chapter 3 (macroscopic modeling), we develop a new protein bead model and employ protein homology technologies for modeling the histone tails of nucleosomes, in the goal of simulating the chromatin fiber using an appropriate macroscopic model and Brownian dynamics. This modeling work on three disparate but complementary scales demonstrates how biomolecular modeling and simulations can be used to generate new insights into the workings of biological systems. |
