Title: "Human Protein Interactome: Reasoning on Structures of Randomness"
Speaker: Jake Chen, Ph.D.; Assistant Professor of Informatics (Indiana University School of Informatics); Assistant Professor of Computer Science (Purdue University School of Science Dept. of Computer and Information Science Indiana University - Purdue University Indianapolis (IUPUI))
Place: Stanley Coulter (SC) 239; Tuesday, September 14, 2004 4:30pm

Abstract

Researchers can gain significant insights into novel protein functions by examining all the interacting (through physical bindings) partners of a protein of interest. The collection of all the interacting proteins for a given organism and the large interaction network that these proteins form is now called the "protein interactome". In the post-genomics era, systematic mapping and mining of protein interactomes for human and model organisms entail findings of new molecular pathways and druggable protein targets. However, high-throughput experimental protein interaction data have previously been believed to be quite noisy and incoherent.

In this talk, I will present major computational opportunities and share biological data analysis insights in our study of ~80,000 proprietary protein interaction data generated from over 200,000 systematic yeast 2-hybrid (SY2H) experiments at Myriad Proteomics. I will begin by briefly describing how we collected, processed, and managed these protein interaction data. Then, I will concentrate my discussions on various computational methods to address both data quality and biologically significant questions, e.g., "How do we assess biases, consistency, and false positive rate in the data", "How do we assess the importance of interactions from replicates", "How do we measure local and global connectivity of a protein in the interaction network", "How do we reason about protein hubs" et cetera. Lastly, I will summarize our experience into lessons for new interdisciplinary researchers who may be interested in advancing this emerging research topic in bioinformatics.


See http://www.stat.purdue.edu/~doerge/BIOINFORM.D/SPRING04/sem.html for a full scheule of BIOINFORMATICS SEMINARS.