Add documentation regarding the assumption of normal distribution for z-scoring #201
Description
Is this a duplicate of an existing idea for this project?
- I found no existing issues covering this topic.
What is your idea?
This idea is about a comment from @MattsonCam in a PR.
When using z-scoring, it assumes the data is in normal distribution. In our case, we can use this method even if it isn't normal distribution and is skewed to one end or another. But, we have found that our method will be less likely to detect poor-quality single cells, especially if the cells are all poor quality and the distribution is more even or Platykurtic.
We should add a description of this assumption, and I have evidence of this limitation we can show too.
Would you like to work on a solution for this?
- Yes I am willing to submit a PR for this!