I am trying calculate Z score of expression from RANseq dataset Ultimate goal is to generate similar figure as in Supplementary Fig 2 B and C of https://www.nature.com/articles/nature24489 I read a discussion (calculate z-score from rpkm values.) that Z score may not be useful in showing the expression changes. Are there any additional steps after simple calculation of z score that may be needed to generate a matrix close to above Fi2 and C above? Any thoughts or suggestions what should be best way to represent change in gene expression using Z score
Gene expression values are generally converted to Z-scores for visualisation purposes, like, in heatmaps. The differential expression analysis itself is performed on normalised counts, prior to log and/or Z-score transformation. Any data can be converted to Z-scores, including FPKM values (edit: see zFPKM R package). The point again, however, is that the Z-score transformation is generally for visualisation in RNA-seq studies, or for other downstream analyses outside of the standard differential expression.
I don't see any major steps that need to be further applied after Z-score transformation for Figure 2B. Figure 2C, however, looks like the plotting of the first 3 dimensions after k-means clustering (hope that we're looking at the same figures).