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!
•
written
by
Sandhiya •
1 answer
The best way to represent count data should be in the original counts,
All that you have FPKM, RPKM, TPM are transformed measures, that unfortunately, may have lost some information along the way.
Different correction factors may have been applied, thus undoing say a TPM is almost never as easy as reverting the formula.
Is there a way where we can convert TPM to raw read count?
If you do wish to give it a go and revert formulas see the timeless post titled:
Log in to answer this question.
-
How to merge heterogeneously normalized RNA-seq datasets for meta-analysis?
written by pipelinfections •
-
Some questions about bulk RNA-seq data normalization
written by feather-W •
-
Convert FPKM to TPM in R
written by AlexStar •
-
Normalizing read counts while having multiple group of samples with different conditions
written by Nemo •
-
Deseq2 - how to save zero counts on normalization step
written by Sandhiya •
-
Need suggestions for RNA-seq quantification tools benchmarking.
written by k.kathirvel93
-
RPKM to TPM conversion?
written by brieannsobieski •
-
differential expression analysis
written by bioinfo456
-
DESeq2 with a small number of genes
written by glihm
-
Large Variation In Rpkm Values Of Different Biological Replicates From Cufflinks
written by hoangtv •
