I took a look at the journal article and it specifies:
Mean coefficient of variation: the mean coefficient of variation across all transcripts.
Source: https://doi.org/10.1093/bioinformatics/bts196
Note that I wasn't able to actually look at the supplementary data since the link in the article never resolved (if someone has the supplementary data, let me know!)
My question is, how is this number calculated? I assume it is calculated from the average coverages of the transcripts. From there I could calculate the standard deviation and average of those numbers and calculate the coefficient of variation that way.
Is this the right approach?