In some cases, like SOX10 (#240) the two distributions: upper / lower overlap significantly.
It would make sense to analyse these genes closer, e.g. check difference with KL divergence.
Theoretically this does not pose a problem: the CDF might jump steep.
Is this "problem" in relation with the Gaussian mixture model?
In some cases, like SOX10 (#240) the two distributions: upper / lower overlap significantly.
It would make sense to analyse these genes closer, e.g. check difference with KL divergence.
Theoretically this does not pose a problem: the CDF might jump steep.
Is this "problem" in relation with the Gaussian mixture model?