HR: 11:55h
AN: NG22A-07 INVITED     [Abstracts]
TI: Searching for Patterns in the Complex World of Protein Interactions
AU: * Samanta, M P
EM: manoj.samanta@systemix.org
AF: NASA Ames Research Center, NASA Genome Research Facility, Moffet Field, CA 94035 United States
AB: Availability of large-scale cellular data from advanced technological approaches has revolutionized biology in recent years, bringing in experts from multiple disciplines. One of the key problems is to determine functions of thousands of novel proteins, whose sequences were determined by the genome-sequencing projects. Data from large-scale protein interaction experiments, conducted to resolve this problem, was noisy and contained many false positives. We developed a graph-theory based statistical approach to correctly predict reliable associations between proteins from noisy interaction data. By further analyzing those associations, we derived tentative functions for 81 unannotated proteins with high certainty. The developed method was robust from the false positives present in the data. In geological sciences, similar inclusion of graph-theory based approaches may help mining for reliable information from noisy large-scale datasets.
UR: http://www.systemix.org/PP/partners/index.php
DE: 0430 Computational methods and data processing
DE: 0466 Modeling
DE: 0520 Data analysis: algorithms and implementation
DE: 0555 Neural networks, fuzzy logic, machine learning
DE: 0599 General or miscellaneous
SC: Nonlinear Geophysics [NG]
MN: Fall Meeting 2005