Date Approved

9-8-2020

Embargo Period

9-9-2020

Document Type

Thesis

Degree Name

M.S. Bioinformatics

Department

Molecular and Cellular Biosciences

College

College of Science & Mathematics

First Advisor

Wu, Chun

Second Advisor

Keck, Thomas

Third Advisor

Chen, Yong

Subject(s)

Computational biochemistry; Drug development

Disciplines

Bioinformatics

Abstract

Markov State Models (MSMs) are constructed from Molecular Dynamics (MD) simulation data, high-resolution spatial and temporal information stored in the form of trajectories, of biological processes, such as ligand-receptor bonding, as a model to understand detailed kinetic information. Traditional MSM implementations involve a clustering step that clusters the MD trajectories into thousands of experimentally unverifiable clusters known as "microstates" before lumping them together into "macrostates". This work details a novel software implementation, using a combination of R, Python, and Tcl, that I have created for the purpose of creating a coarse-grained MSM that directly clusters the MD trajectories into a handful of experimentally verifiable clusters while maintaining the Markovian property. The coarse-grained MSM implementation was designed to require minimal technical experience while still being robust enough for usage in studying a variety of biological processes. In addition, this coarse-grained MSM implementation has already been used as part of several works to explore the binding mechanisms of various ligand-receptor complexes that have shown potential in the treatment of neurodegenerative diseases and various cancers.

Share

COinS