I am analyzing the TCGA data and doing the differential expression analysis for about 6,000 samples and 20,000 coding genes.
I was suggested to use DESeq2 for DE analysis, but the DESeq() takes an extremely long time with 6,000 samples, therefore I would like to use the limma-voom instead. In the tximport vignette, the scaled counts generated from abundance ("scaledTPM" or "lengthScaledTPM") are recommended to be used in limma-voom. However, I only have the tximport object generated from Salmon and its countsFromAbundance = "no". Thus in this case txi$abundance and txi$counts are the gene-level summarized TPM table and count table from the original Salmon .sf files.
My question is, can I use the gene-level TPM or count table to get counts from abundance? I have no access to the fastq files, and TPM can't be converted into CPM. Or any suggestions for other tools? Thanks!