Came across your paper and found your tool to be useful. I especially want to implement the CV score to compare batch correction methods and different iterations for my own data that are naturally created during any analysis.
I was wondering why the cv scores are not normalized when comparing between methods? Is it because all scores are calculated by normalizing over the mean (this should be for gene to gene comparisons)? I wanted to know your thoughts on log transforming then normalizing before comparing between two different batch correction methods. I think this would remove the magnitude of variation that the values can convey, but limit values between 0-1....appreciate your thoughts.
Hello,
Came across your paper and found your tool to be useful. I especially want to implement the CV score to compare batch correction methods and different iterations for my own data that are naturally created during any analysis.
I was wondering why the cv scores are not normalized when comparing between methods? Is it because all scores are calculated by normalizing over the mean (this should be for gene to gene comparisons)? I wanted to know your thoughts on log transforming then normalizing before comparing between two different batch correction methods. I think this would remove the magnitude of variation that the values can convey, but limit values between 0-1....appreciate your thoughts.
Thanks, Shahroze