This preview shows pages 1–2. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.View Full Document
Unformatted text preview: Vol. 23 ISMB/ECCB 2007, pages i222i229 BIOINFORMATICS doi:10.1093/bioinformatics/btm222 Systematic discovery of functional modules and context-specific functional annotation of human genome Yu Huang 1, , Haifeng Li 1, , Haiyan Hu 1 , Xifeng Yan 2 , Michael S. Waterman 1 , Haiyan Huang 3 and Xianghong Jasmine Zhou 1, * 1 Molecuolar and Computational Biology, University of Southern California, Los Angeles and 2 IBM T. J. Watson Research Center, Hawthorne, NY and 3 Department of Statistics, University of California, Berkeley, CA, USA ABSTRACT Motivation: The rapid accumulation of microarray datasets provides unique opportunities to perform systematic functional characteriza- tion of the human genome. We designed a graph-based approach to integrate cross-platform microarray data, and extract recurrent expression patterns. A series of microarray datasets can be modeled as a series of co-expression networks, in which we search for frequently occurring network patterns. The integrative approach provides three major advantages over the commonly used micro- array analysis methods: (1) enhance signal to noise separation (2) identify functionally related genes without co-expression and (3) provide a way to predict gene functions in a context-specific way. Results: We integrate 65 human microarray datasets, comprising 1105 experiments and over 11 million expression measurements. We develop a data mining procedure based on frequent itemset mining and biclustering to systematically discover network patterns that recur in at least five datasets. This resulted in 143 401 potential functional modules. Subsequently, we design a network topology statistic based on graph random walk that effectively captures characteristics of a genes local functional environment. Function annotations based on this statistic are then subject to the assessment using the random forest method, combining six other attributes of the network modules. We assign 1126 functions to 895 genes, 779 known and 116 unknown, with a validation accuracy of 70%. Among our assignments, 20% genes are assigned with multiple functions based on different network environments. Availability: http://zhoulab.usc.edu/ContextAnnotation Contact: email@example.com 1 INTRODUCTION Systematic functional characterization of genes identified in the genome sequencing projects is urgently needed in the post- genomic era. The rapid increase in large-scale gene expression data provides us unique opportunities to meet this need. A commonly used approach is to cluster genes with similar expression patterns (Beer and Tavazoie, 2004; Gasch and Eisen, 2002; Tamayo et al. , 1999), and to predict functions of unknown genes based on their expression similarity to known genes (Gasch and Eisen, 2002; Niehrs and Pollet, 1999)....
View Full Document
- Spring '09