Here is a new tool for Gene Ranking and differential expression analysis of RNA-Seq Data via Discriminant Non-Negative Matrix Factorization (DNMF).

1. What can DNMF do?

Gene ranks of RNA-seq for gene set enrichment analysis (GSEA) due to there are no built-in ranking methods for RNA-seq data. Or other tools requiring gene ranks. Additionally, differential expression analysis of RNA-Seq Data

2. Why DNMF?

Please read Jia Z, Zhang X, Guan N, Bo X, Barnes MR, Luo Z (2015) Gene Ranking of RNA-Seq Data via Discriminant Non-Negative Matrix Factorization. PLoS ONE 10(9): e0137782. doi:10.1371/journal.pone.0137782

3. How it works?

There are two types of samples, A and B, in the matrix *V*, each with four replicates. Via DNMF, two matrices, *W*and *H*, are produced. The matrix *H* possesses the same column number as matrix *V*, while the matrix *W*possesses the same row number as matrix *V*. To explain it biologically, the row of *H* could be viewed as the metagenes (the up-regulated or down-regulated gene sets), while the column of *W* represents the weight of each gene in the metagenes. The two metagenes in matrix *H*, represent the up-regulated genes and down-regulated genes respectively. Due to the correspondence between columns of *W* and rows of *H*, it is possible to identify which metagenes are combined with the up-regulated or down-regulated gene sets in matrix *W*. Consequently, it is easy to obtain the rank of genes by subtraction between the weights of the two metagenes in matrix *W*.

4. Code and bug reports

https://github.com/zhilongjia/DNMF

https://cran.r-project.org/web/packages/DNMF/index.html

Thank you.