nonnegative matrix factorization an analytical and interpretive tool in computational biology非负矩阵分解的分析和计算生物学的解释工具.pdfVIP

Review Nonnegative Matrix Factorization: An Analytical and Interpretive Tool in Computational Biology Karthik Devarajan* Division of Population Science, Fox Chase Cancer Center, Philadelphia, Pennsylvania, United States of America sample to predict its class. Once a list of potentially interesting Abstract: In the last decade, advances in high-through- genes has been identified from these analyses, one is often put technologies such as DNA microarrays have made it interested in characterizing these genes in terms of function. In this possible to simultaneously measure the expression levels paper, we review nonnegative matrix factorization (NMF) and its of tens of thousands of genes and proteins. This has applications in computational biology, with an emphasis on the resulted in large amounts of biological data requiring analysis and interpretation of high-throughput biological data such analysis and interpretation. Nonnegative matrix factoriza- as those above. We discuss and illustrate the properties of NMF tion (NMF) was introduced as an unsupervised, parts- based learning paradigm involving the decomposition of through examples from the literature, and provide an intuitive a nonnegative matrix V into two nonnegative matrices, W interpretation of the factorization and its implicit sparse nature as and H, via a multiplicative updates algorithm. In the well as the nonnegativity constraints. In particular,