[Todos] Tesis de Licenciatura de Patricio Traverso /Area de interes: Bioinformatica

[Igor Zwir WashU]

Presentación de Tesis de Licenciatura en Ciencias de la Computación

       "Learning Robust Dynamics in Bacterial 
          Regulatory Networks Using GENIE"


Alumno:   Patricio Traverso
Director:   Igor Zwir

Jurados:   Pablo Jacovkis,  Alberto Kornblihtt , Irene Loiseau

Jueves 2 de septiembre, 15:30 hs, aula a confirmar.

Están todos invitados.


ABSTRACT:

One of the big challenges of the post genomic era is determining
when, where and for how long genes are turned on or off. Gene
expression is determined by protein-protein interactions among
regulatory proteins and with RNA polymerase(s), and protein-DNA
interactions of these trans-acting factors with cis-acting DNA
sequences in the promoters of regulated genes. These interactions
define complex genetic networks, which designs have motivated
researchers to draw direct analogies with established techniques
in electrical engineering. As with the construction of electrical
circuits, the gene circuit approach uses mathematical and
computational tools in the construction and posterior analysis of
a proposed network diagram. So far, the qualitative agreement
between model and experiment in a series of studies depends both
on the design of the network topology, which most of the times
includes uncertain connections between genes, as well as on the
dynamic behavior of the network, which is affected by the
ambiguity inherent to the biological processes (e.g., monomer or
dimer binding of promoters, enzymes having kinase and/or
phosphatase activities, etc.) and the mathematical models used to
represent them (e.g., Boolean or continuous models; reverse or
forward algorithms).  Moreover, the number of genes considered in
the networks is usually large compared to the number of the
available measurements (e.g., time-point expression), thus, more
than one possible model may be consistent with the subjacent data.
Finally, the data always contains a substantial amount of noise,
provided by the systematic variability of the experiments, which
in addition to previous problems, makes it difficult to deduce the
implications  of the underlying logic of genetic networks through
experimental techniques alone.

We propose a methodology, termed GENIE for Gene Expression
Networks Iterative Explorer, that embraces the uncertainty
inherent to the biological problem and the imprecision of their
underlined mathematical models by using an iterative approach.
First, GENIE proposes a network topology based on DNA sequence
analysis of transcription factor interactions, which, together
with previous knowledge from the literature, constitute the raw
material for the architecture design.  Second, we transform the
hypothesis provided by the network topology, by means of its
possible chemical reactions and physical constraints, into a
system of nonlinear ordinary differential equations, whose
variables are concentrations of proteins, mRNA, etc.  Rather than
advocating a single, definitive model of the genetic network, we
describe a variety of models that have different strengths,
weaknesses and domains of applicability. Third, the network models
are evaluated by testing the ability to reproduce the behavior
observed in vivo of their subjacent non-linear models, each
characterizing the time-dependent change in concentration of the
components, including kinase, phosphatase, and transcription
activities. Fourth, the successful models are tested by
considering different emergent systemic properties, such as
flexibility to reproduce all possible functional patterns, and
robustness to changes in parameters and initial conditions. Fifth,
we revisit the original topology and iterate, developing adaptive
models of genetic networks.  Finally, a decision making process
reveals the most realistic models, which are examined by a
datamining process providing substantial insights from the modeled
genetic systems.

We applied GENIE to uncover regulatory networks in the enteric
bacteria Salmonella enterica serovar Typhimurium by focusing on
the PhoP/PhoQ and PmrA/PmrB two-component systems, which govern
virulence and the adaptation to low Mg2+ and high Fe3+
environments, respectively.  The study of the PhoP regulon
constitutes a special challenge due to the multiplicity of
PhoP-controlled targets, and the connectivity of the PhoP/PhoQ
system with other two-component systems, such as PmrA/PmrB,
transcriptional regulators, and alternative RNA polymerase sigma
factors.  We verified our predictions by investigating
transcription and PhoP and PmrA protein binding to the identified
promoters in vivo.



Igor Zwir
Howard Hughes Medical Institute
Department of Molecular Microbiology, Box 8230
Washington University School of Medicine
St. Louis, MO 63110-1093 USA
Phone, lab: (314)-362-3691
FAX: (314)-747-8228
email: zwir at borcim.wustl.edu
http://www.microbiology.wustl.edu/dept/postdoc/zwir
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.de.fcen.uba.ar/pipermail/todos/attachments/20040825/f1afcb62/attachment.htm