I have a big table, its rows are genomic coordinates and columns are the genomic features (like below). I would like to separate rows and columns based on the variability, I have tried to use some basic statistics like below codes, but I like to know is it the right way or is there an alternative (statistical) way that would be more accurate?
Feature_A Feature_B Feature_C Feature_D cord_1 0.9 1 0.8 1 cord_2 0.6 0.1 0.9 0.5 cord_3 0 0 0 0 cord_4 0.1 0 0 0.2
DF$skew<-rowSkewness(DF) DF$var <-rowVars(DF) DF$sd <-rowSds(DF) DF$IQR <- rowIQRs(DF)) DF$mean <- rowMeans(DF) DF$coef.var <- DF$sd /DF$mean
I would like to consider cord_2 (as more variable) and ignore cord_1,3 and 4 in my output, so based on that, which statistic element is more better?