Arabidopsis thaliana gene transcript modeling with a custom Bayesian-based CHC genetic algorithm
Bel LaPointe and David John
Department of Computer Science, Wake Forest University
Friday, November 17, 2017, 12:00 – 1:00 pm
Gene interaction models are created using multiple sets of experimentally collected Arabidopsis time course gene transcripts. This recent work is founded on previous ideas (e.g. Bayesian statistics) where a Metropolis-Hastings algorithm was used for this purpose. Unfortunately, the Metropolis-Hastings approach will not scale to larger data sets. A new approach, based on a CHC Genetic Algorithm, has been developed that does scale to larger data sets. As well, this new algorithmic approach allows for non-uniform priors and “black listing”. An explanation of how this new genetic algorithm works as well as presentations on the application of this algorithm to several Arabidopsis data sets will be highlighted.
Categories: past events