Normalizing read counts while having multiple group of samples with different conditions

Question

[Nemo]

Hello, I have a group of samples which have different conditions. It means for example 10 of them have disease A and 10 disease B and the other 10 disease C and 10 more normal samples. All these diseases are in relation to each other. I have read counts after alignment and I would like to normalize them. When I look into FPKM, RPKM and TPM, all have the common feature that can be applied on a group of samples from the same group. My question is how I can normalize them?

I just read about TMM, but I do not know exactly if it is suitable? and how I can implement it on my data.

Answer 1

[seidel]

The TMM algorithm is designed to give good normalization across differing conditions, and is incorporated into the edgeR library in R. There's a good discussion and some code for how to get TPMs from normalized read counts discussed by Gordon Smyth in a previous thread. The edgeR userguide is always a good read.