Filling gaps in a metabolic network using expression information

Authors:
Peter Kharchenko;Dennis Vitkup;George M. Church
Affiliations:
Department of Genetics, Harvard Medical School, 77 Louis Pasteur Avenue Boston, MA 02115, USA;Department of Genetics, Harvard Medical School, 77 Louis Pasteur Avenue Boston, MA 02115, USA;Department of Genetics, Harvard Medical School, 77 Louis Pasteur Avenue Boston, MA 02115, USA
Venue:
Bioinformatics
Year:
2004

Citing 0
Cited 7

Comparison of reversible-jump Markov-chain-Monte-Carlo learning approach with other methods for missing enzyme identification

Journal of Biomedical Informatics
SWARM: a scientific workflow for supporting bayesian approaches to improve metabolic models

CLADE '08 Proceedings of the 6th international workshop on Challenges of large applications in distributed environments
Max-margin Classification of Data with Absent Features

The Journal of Machine Learning Research
Comparison of Bayesian and regression models in missing enzyme identification

International Journal of Bioinformatics Research and Applications
A Bayesian Approach to High-Throughput Biological Model Generation

BICoB '09 Proceedings of the 1st International Conference on Bioinformatics and Computational Biology
Transcription factor centric discovery of regulatory elements in mammalian genomes using alignment-independent conservation maps

RCG'06 Proceedings of the RECOMB 2006 international conference on Comparative Genomics
An evolutionary approach for searching metabolic pathways

Computers in Biology and Medicine

Quantified Score

Hi-index	3.84

Visualization

Abstract

Motivation: The metabolic models of both newly sequenced and well-studied organisms contain reactions for which the enzymes have not been identified yet. We present a computational approach for identifying genes encoding such missing metabolic enzymes in a partially reconstructed metabolic network. Results: The metabolic expression placement (MEP) method relies on the coexpression properties of the metabolic network and is complementary to the sequence homology and genome context methods that are currently being used to identify missing metabolic genes. The MEP algorithm predicts over 20% of all known Saccharomyces cerevisiae metabolic enzyme-encoding genes within the top 50 out of 5594 candidates for their enzymatic function, and 70% of metabolic genes whose expression level has been significantly perturbed across the conditions of the expression dataset used. Availability: Freely available (in Supplementary information). Supplementary information: Available at the following URL http://arep.med.harvard.edu/kharchenko/mep/supplements.html