Automated Gene Classification using Nonnegative Matrix Factorization on Biomedical Literature

Understanding functional gene relationships is a challeng ing problem for biological applications. High-throughput technologies such as DNA micr oar ays have inundated biologists with a wealth of information, however, processing tha t information remains problematic. To help with this problem, researchers have begun appl ying text mining techniques to the biological literature. This work extends previous wo rk based on Latent Semantic Indexing (LSI) by examining Nonnegative Matrix Factorizat on (NMF). Whereas LSI incorporates the singular value decomposition (SVD) to appro ximate data in a dense, mixedsign space, NMF produces a parts-based factorization that i s directly interpretable. This space can, in theory, be used to augment existing ontologies and annotations by identifying themes within the literature. Of course, performing NMF doe s not come without a price— namely, the large number of parameters. This work attempts t o analyze the effects of some of the NMF parameters on both convergence and labeling accur y. Since there is a dearth of automated label evaluation techniques as well as “gold st andard” hierarchies, a method to produce “correct” trees is proposed as well as a technique to label trees and to evaluate those labels.