Graduate Thesis Defense - 2015 Spring

Speaker: Geoffrey Lentner, University of Louisville

Abstract: The prospect of detecting an Extraterrestrial Intelligence (ETI) is supreme and its implications for our collective human perspective goes without saying. There are many ways in which an estimate on the total number of ETIs in the Milky Way can be expressed, the most prominent construction being the Drake Equation. In a new approach, I turn this endeavor on it’s head and instead ask the question, given a solution N, where might me expect to find our nearest neighbor? As we compile more data with projects such as SDSS-III we will be able to model the spatial distribution of key parameters in our search for life and habitability in the Milky Way.

I will present a code I’ve developed for the purpose of numerically modeling galaxies via user defined statistics. Generally speaking, this software has the capability to generate any system of particles confined to a 3D volume. The user defines an arbitrary number of probability density functions (PDFs) in any of the three major coordinate systems. The additional step this program makes is to build an entire ensemble of these systems and perform a nearest neighbor analysis on each to construct functions that describe the expectation of finding a neighbor inside a certain volume given the location of inquiry. There is the potential here for impact with respect to the Fermi paradox and how preferable a location we live in to detect ETIs and/or find habitable worlds.

Speaker: Michael Martin, University of Louisville (PhD Candidacy)

Abstract: Computer models of life forms are quantitatively predictive of metabolic parameters throughout an organism’s entire life cycle [1]. While to date, whole cell models mostly include rudimentary bacteria with very small genomes (525 genes), the field is rapidly moving to more sophisticated eukaryotic systems (20,000 genes in the human genome). Eventually models of human cells will help to elucidate cancer mechanisms, differentiation of stem cells into somatic cells, and the process of aging. The gaps in our mechanistic understanding of the molecular interactions within cells are being filled by higher fidelity computational cell models.

In this presentation, the basics of computational cell biology are presented along with a summary of key physical processes within the cell. Of relevance is the fact that protein-protein and enzyme-substrate interactions may be mathematically represented via the “Law of Mass Action” to construct sets of linear differential equations that describe reaction rates and formation of protein complexes. Using mass action methods, examples of reaction networks are presented within the Virtual Cell simulation package.

The proposed thesis topic focuses on computational modeling efforts to capture the epigenetic regulation of an ancient (typically dormant) retro-viral genetic element, named LINE-1 (long interspersed nuclear element 1). This remnant of viral infection has managed to survive within the human genome and copied itself thousands of times to now represent 17% of human DNA – part of what was once labelled junk DNA. When activated, the proteins encoded by LINE-1 function to produce copies of itself (along with other similar genetic elements) that can then be reinserted into the genome. Thus, activation of LINE-1 is associated with genomic instability and tumorigenesis. Initial results are presented for simulation of the retinoblastoma pathway and its association with the activation of LINE-1.

References:
1. Jonathan R. Karr, Jayodita C. Sanghvi, Derek N. Macklin, Miriam V. Gutschow, Jared M. Jacobs, Benjamin Bolival, Nacyra Assad-Garcia, John I. Glass, Markus W. Covert. A Whole-Cell Computational Model Predicts Phenotype from Genotype. Cell, 2012; 150 (2): 389 DOI: 10.1016/j.cell.2012.05.044