Abstract
Applied to microarray data, Nonnegative Matrix Factorization (NMF) can be viewed as a generalized clustering algorithm allowing for gene overlaps - an important feature in this domain where genes can be involved in several biological processes. In this paper we present siNMF, a generalization of NMF that can simultaneously factorize a gene expression matrix and a matrix of transcription regulatory influences. Thus, siNMF constructs gene clusters taking into account not just expression information, but also background knowledge on potential regulatory factors of the clusters. A preliminary application of the algorithm to a real-life pancreatic cancer dataset shows the feasibility of our approach.

URL:
http://rewerse.net/publications/rewerse-publications.html#REWERSE-RP-2007-027

BibTeX:

@inproceedings{REWERSE-RP-2007-027,
	author = {Liviu Badea},
	title = {Combining Gene Expression and Transcription Factor Regulation Data using Simultaneous Nonnegative Matrix Factorization},
	booktitle = {Proceedings of 2007 International Conference on Bioinformatics and Computational Biology, Las Vegas, Nevada, USA (25th--28th June 2007)},
	year = {2007},
	pages = {127--131},
	url = {http://rewerse.net/publications/rewerse-publications.html#REWERSE-RP-2007-027}
}