What is the meaning of the sample view curve on the normalization step? does it need to be like on a normal distribution as well?
This can help you find outliers that should be removed. If you see one/several samples are extremely different than the others, it should be excluded as there was probably a technical error with this sample and it could greatly influence the downstream statistical analysis.
Thank you for taking your time to answer my question.
However I can’t seem to understand what you mean. Could you please be more clear.
Thank you in advance!
We don’t care whether the sample-wise distribution is approximately normal, since all statistics are done feature-wise.
However, sample-wise distribution can still be helpful for QA/QC purposes. If one or several sample-wise distributions are extremely different from the majority, it could indicate that there were technical problems with the data collection and those samples should be excluded from the analysis.