Keywords: regulation, sequence, alignment, mycobacteria tuberculosis,
Bayesian algorithms |
Over the course of its life cycle, mycobacteria turbercolisis
exists in a number of distinct environments, including at least three distinct
hostile environments inside human macrophage cells. Differential gene expression
plays a major role in it adaptation to the distinct environments. Accordingly,
it has a particularly complex system of transcription regulation compared
to other prokaryotes. My research will focus on understanding transcription
regulation using an approach based in computational statistics. If specific
approaches can be developed, computational analyses promise to accelerate
the identification of regulatory regions in genomic sequences at minimal
cost. It appears that cross-species comparison of genomic sequences, phylogenetic
footprinting, offers great promise to overcome the specificity challenge.
In response, my research will focus on the computational characterization
of gene regulatory modules through the development and application of advanced
Bayesian alignment algorithms. To date only very limited progress has been
made in this area, but Dr. Lawrence and his collaborators have recently
shown that Bayesian alignment methods are very well suited to phylogenetic
footprinting. However, the first Bayes alignment algorithm also has limitations,
assuming that conserved regions align perfectly, with no gaps. In many
cases, it would not be desirable to permit small gaps within these conserved
regions, but still ignore unrelated portions of the sequences. This has
the distinct advantage of using Bayesian inference methods to find the
distribution of an alignment plus the distribution within each of these
conserved regions, giving all posterior inferences for all variables including
the number of local alignments. Once my research is completed, the new
algorithm then will yield the posterior distribution of locally aligned
conserved regions of a pair of sequences, or local subsets, in the multiple
sequence case. While we expect the algorithms that I develop to have other
applications, my work will focus on gene expression in mycobacteria turbercolisis
(MTB). |
