Entering edit mode
3.8 years ago
taylor
•
0
Recently, I got two tables of TPM values from my colleague, and the raw sequencing data were missing... I have no idea how to use these data to identify different expressed genes. Since I always analyze different expressed gene by DEseq2, I am afraid it cannot help me on this situation. So is there any software recommended? Thanks~
Well, if you neither have raw reads nor raw counts, your starting point are these TPM values. But if you have raw counts you should start with that. You cannot go from TPM to raw count because you normalize by the number of reads, which is an information that you do not have.
You can still use your TPM tables to do some visualization like heatmap : RNAseq TPM heatmap
Thanks, but I am afraid I have to get a list of different expressed genes... And what do you think about the results I got from DESeq/Limma by using TPM? Will that still make a little sense?
You can take a look at Gordon' point of view about DE analysis starting fromTPM here : https://support.bioconductor.org/p/98820/
You should get your collegues to give you the count data. Whichever tool calculated the TPM values most likely also calculated the (estimated) counts. Also please note if you don't have the raw data you cannot publish as that is a requirement for reproducibility.
Ah yes, we know that... We just want to see if there is any evidence before we redo those experiments. However, we do not have count data but TPM now...
I'd recommend viewing at the Biostars post targeting a similar question: How to use TPM from RNA seq data analysis for differential gene expression analysis? which statistical methods are reuired to be performed.