Hi all,

I'm trying to collect the datasets from public datasets from GEO. I found some data in fpkm and some in rpkm,tpm. My goal is to feed these data into a machine learning model, but my question is "ALL THESE DATA SHOULD BE IN A COMMON FORMAT "; to do this what's the best solution?

I have two questions:

1) To have a common unit, I should convert these data to tpm ie., fpkm,rpkm to tpm; so that all values are in same format In this case how can I do this conversion?

2) If I decide to use DESeq2 to normalize; then is it right to feed all these different file formats into DESeq2 and uniformly have deseq2 normalized values?

Please help me with this?

Thanks!

You won't be able to make a DESeqDataSet without the count data. If you just have FPKM/TPM this is going to be hard for you.

You can convert all to TPM and then consider other tools for batch correction and normalization.