I have more than 100 RNA-Seq samples (files) from different labs (i.e, different papers) for plant tissues. I want to compare them (for DEGs, PCA, tSNE, absolute expression, etc.). But for this, the data should be normalized. Data are from different labs, thus I want to take special care for the normalization which can remove any bias. I know the bias may depends on many factors. But here as a first step, I want to know what kind of normalization should be applied to the single count data (matrix)? I guess TMM is one of the best normalization technique in this scenario. Is there any other pipeline for this kind of task? and lastly, can I just compare the data with TPM values in doing PCA and tSNE?
Thanks for help!