Ph.D. thesis, Travis Hoppe
Full LaTeX source and final product.
Type make from the command-line to build the thesis.
Assuming you have a standard build this should be fine, though I pull heavily from the "extras" packages in the Ubuntu builds.
To view the final PDF without cloning the repo, click "save-as" in on the link and view from your local PDF viewer.
The final thesis is about 150 pages in the "library-format" double spacing, but realistically is only about 100 pages using single spacing. Your committee members won't be impressed by the pagecount, give them a single spaced double-sided printout (unless they ask otherwise!). Remember that your thesis only needs to be "good enough", so finish it up and get back to science!
A protein's ultimate function and activity is determined by the unique three-dimensional structure taken by the folding process. Protein malfunction due to misfolding is the culprit of many clinical disorders, such as abnormal protein aggregations. This leads to neurodegenerative disorders like Huntington's and Alzheimer's disease. We focus on a subset of the folding problem, exploring the role and effects of entropy on the process of protein folding. Four major concepts and models are developed and each pertains to a specific aspect of the folding process: entropic forces, conformational states under crowding, aggregation, and macrostate kinetics from microstate trajectories.
The exclusive focus on entropy is well-suited for crowding studies, as many interactions are non-specific. We show how a stabilizing entropic force can arise purely from the motion of crowders in solution. In addition we are able to make a a quantitative prediction of the crowding effect with an implicit crowding approximation using an aspherical scaled-particle theory.
In order to investigate the effects of aggregation, we derive a new operator expansion method to solve the Ising/Potts model with external fields over an arbitrary graph. Here the external fields are representative of the entropic forces. We show that this method reduces the problem of calculating the partition function to the solution of recursion relations.
Many of the methods employed are coarse-grained approximations. As such, it is useful to have a viable method for extracting macrostate information from time series data. We develop a method to cluster the microstates into physically meaningful macrostates by grouping similar relaxation times from a transition matrix.