Applied Mathematics Seminar
- Thursday, November 12, 2015 from 3:00pm to 4:00pm
- Wilson Hall - view map
On Thursday Novermber 12th, Bree Cummins from Department of Mathematical Sciences will present an Applied Math seminar"Matching gene transcription data with proposed mechanisms of gene regulation"Abstract:We are using switching models of gene regulatory networks to propose mechanisms of gene regulation given time series data. Often, little to nothing is known about the interactions between gene products and DNA. Moreover, the number of experiments needed to discover interactions grows combinatorially with the number of genes. The purpose of our research is to narrow the potential number of experiments by identifying networks that can reproduce features of the time series data. Switching models have the property that they are analytically solvable, and so the exploration of parameter space is relatively cheap. Our collaborators have written a software tool called Dynamic Signatures of Gene Regulatory Networks (DSGRN) that can analyze the switching model of a gene regulatory network over all of parameter space. The output is a coarse but global characterization of the dynamics, from which we can identify the presence and general location of stable fixed points and stable and unstable periodic orbits. Often, the coarse dynamics of gene regulation can be identified from the data, and so desirable parameterizations of a model can be selected for further examination. I will briefly introduce these ideas in more detail, and then describe the process of taking output from DSGRN and matching it against the original time series data to discover which parameterizations are reasonable matches for the experimental observations. I will frame this task in terms of matching a partial order against a labeled graph, and I will demonstrate using data from the malarial parasite Plasmodium falciparum.