Entering edit mode
8.0 years ago
mariamari693693
▴
50
I have two RNA-seq datasets. Both of them have been prepared in the same way. I have only normalized count data for both. I thought to simply merge and normalize them again (which means normalization of normalized values) and then performing a batch effect. Does anybody have any idea to share with me? what would be the potential risks that I did not considered
How were the data normalized? If they are FPKM you probably don't need to repeat the normalization (actually, you shouldn't). If they were normalized in some other way, then it depends. However, if I well remember normalization methods used (e.g.) by DESEQ require raw counts and do not expect to have normalized data as input, but I didn't investigate the issue and I am not sure if you can do normalization of a normalization.
Dear Mariam, Hi
If you have performed in silico normalization such as what Trinity will do (in the last version, it is by default), I guess it is not a good idea to launch normalization, again.
~ Best