Entering edit mode
8.3 years ago
Xiaokang
▴
80
I'm using the RNA-Seq data from TCGA of Level 3 with gene expression levels of normal and tumor patients to do classification. But found that, even in the RSEM_genes_normalized file, there are so many genes with all-zeros across all samples, or with many-zeros across samples. Since for the all-zeros genes, it seems useless, so I remove those genes. But for the many-zeros genes, how should I deal with them? Are the zeros missing values that need imputation?
