"Expression Profiling Efficiency" description in Metrics.md

[paulimer]

Hi and thanks for the revival of this very useful tool! All other RNA-seq tools are quite difficult to use, with specialized inputs, so it is very helpful to have one so straightforward.

I had a question about your description of the Expression Profiling Efficiency. In the Metrics.md file, it is indicated

The proportion of "Exonic Reads" (see "Exonic Rate", below) out of all reads which were not Secondary Alignments or Platform/Vendor QC Failing reads.

This would mean that it is something like "Exonic Reads / Mapping reads". But this would be the same as the exonic rate.

Seeing that the Expression Profiling Efficiency is actually "Mapping Rate * Exonic Rate", it looks like it is more akin to "Exonic Reads / Total Reads", maybe a rewording of the description might be useful. When using MultiQC, the popup help actually says "Ratio of exonic reads to total reads".

Or maybe it is only I that is confused by the wording of the description!

[Rohit-Satyam]

I am confused how the % Expression Efficiency column in multiqc is being calculated. When I hover on the header it says that this percentage is ratio of exonic reads to total reads. To check that, I referred to Total Reads 18128 and Exonic Reads 2916 values from metrics.tsv file. This math gives me roughly 16% efficiency however, my multiqc reports 61.9%.

This proportion is also not equal to Expression Profiling Efficiency 0.398906 or 1-Expression Profiling Efficiency which will be 60.12%

[agraubert]

The difference between Mapped Reads and Unique Mapping, Vendor QC Passed Reads is that the latter doesn't actually need the read to be mapped. Unique Mapping, Vendor QC Passed Reads just makes sure that none of the Secondary, QCFail, or Supplementary flags have been set