Date |
April 23, 2007 |
Speaker |
Dr. Yang Dai, Department of Bioengineering, University of Illinois at Chicago |
Title |
Reconstruction of Functional Gene Networks based on Published Literature |
Abstract |
Genomics and proteomics technologies have provided unprecedented
opportunity to providing a comprehensive survey of the cell's
interactome. The mapping of interactions among genes/proteins is
extremely important for the understanding of underlying mechanism of
the cell. These technologies provide unbiased opportunity for the
study.
However, large body of false positive interactions also presents
problems. These problems are partially alleviated by performing (1)
the rigorous statistical analysis to derive significant results, and
(2) the integration of various genomic, proteomic data and prior
biological knowledge. The utilization of prior biological knowledge,
especially the information from published literature, can be extremely
important.
However, the existing methods are still primitive; basically
forming the networks by taking the sum of all interactions identified
from published literature. The resulted networks are static and do
not provide information on interactions that have yet to be
discovered. Systematic analyses of the literature networks and
extraction of knowledge that can be used for the assessment of
biological relevance of detected interactions from the experimental
data are clearly needed. In this study we address this issue. We
propose a novel statistical method to compute the likelihood of
interaction (LOI) scores for interactions of each type (regulation,
expression, binding, inhibition, and protein modification) obtained
from the databases of published literature.
The Gene Ontology Molecular Function annotations will be used for
the representation of functional resemblance among interactions. We
demonstrate the effectiveness of this approach through the
validation of published literature on yeast and human at genome scale
and as well as KEGG and MARK-Signaling curated yeast pathways.
|
|