MSNet

Mining gene functional networks to improve mass-spectrometry based protein identification

Smriti R. Ramakrishnan[1], Christine Vogel[2], Taejoon Kwon[2], Luiz O. Penalva [3], Edward M. Marcotte[2], Daniel P. Miranker[1]


[1] Department of Computer Sciences, 1 University Station C0500, The University of Texas at Austin, Austin, TX 78712
[2] Center for Systems and Synthetic Biology, Department of Chemistry and Biochemistry & Institute for Cellular and Molecular Biology, 2500 Speedway, The University of Texas at Austin, Austin, TX 78712

[3] Children's Cancer Research Institute; The University of Texas Health Science Center at San Antonio; San Antonio, TX 78229

Contact: DPM (miranker at cs utexas edu) or EMM (marcotte at icmb utexas edu)

This website provides software and supplementary data used in MSNet analyses.

Abstract

Tandem mass spectrometry (MS/MS) promises fast and reliable characterization of complex protein mixtures, but suffers from low sensitivity in protein identification. In a typical shotgun-proteomics experiment, it is assumed that all proteins are equally likely to be present. However, there is often other evidence to suggest that a protein is present and confidence in individual protein identification can be updated accordingly. We develop a method that places an MS/MS experiment in the larger biological context of pathways and functional associations between proteins. Our method integrates a gene functional network with the proteins identified in MS/MS experiments. As a result we can substantially increase in the number of proteins identified at the same error rate. We predict up to 10-22% more proteins than the original MS experiment when applied to yeast in different experimental conditions analyzed on various MS/MS instruments, and up to 48% more proteins on a human sample, validating up to 96% of our yeast predictions by presence in ground-truth reference sets.

Publications

Resources

Last modified: smriti, July 2009