search

cbielow / PTXQC

A Quality Control (QC) pipeline for Proteomics (PTX) results generated by MaxQuant
Other
42 stars 25 forks source link

Labeling efficiency #4

Closed cbielow closed 6 years ago

cbielow commented 8 years ago

In OpenMS the ITRAQAnalyzer will report the labeling efficiency for iTRAQ channels, i.e. how many spectra contain a certain reporter ion. High values indicate successful chemical labeling, high amount of sample and few contamiants (usually not labeled since introduced post labeling).

Low values indicate that either of the above is not optimal, or that a large portion of proteins is not expressed in certain samples, but in others, e.g. if you mix yeast (labeled, 114, 115) + HEK (label 116, 117), then HEK peptides will lack the 114,115 label, whereas yeast peptides lack the 116,117 label. Since HEK has more proteins/peptides, labeling efficiency in 114,115 will be lower, since we average over all MS/MS spectra.

The MQ APK files will probably suffice as the data source, but it might be time-consuming to parse them. Maybe we need to write a C(++) parser, since R is really not meant for this....

This also works for TMT of course.

cbielow commented 8 years ago

this can (n theory) also be done using the 'reporter.XX.corrected' columns in msms.txt, but as of now, MaxQuant does not fill them (at least not when using MS3 multi-notch) -- issue is currently reported at MQ tracker... lets see if this is fixed soon -- then we can implement this very valuable metric.

cbielow commented 8 years ago

MaxQuant bug: http://maxquant.myjetbrains.com/youtrack/issue/MaxQuant-150

TMT reporter intensity in msms.txt
I'd like to extract the labeling efficiency, i.e. fraction of TMT-labeled peptides, which will help during method optimization.
Sometimes, many peptides remain unlabeled and can therefore not be identified (since they are missing the TMT modification which is set as fixed).
The columns 'reporter.xx.' are present in msms.txt, but remain empty when using MS3 multi-notch. This is somehow understandable, since reporters are quantified in MS3, but the ms3Scans.txt does not have them either. 

So: is it possible to fill the 'reporter.xx' columns in msms.txt?! If need be also ms3Scans.txt is fine, but having one place (for multi-notch vs. normal MS2-HCD) would be preferrable.

MaxQuant version used: 1.5.3.30
cbielow commented 6 years ago

discussed on MQ mailing list here: https://groups.google.com/forum/#!topic/maxquant-list/m352CHcA81g%5B1-25%5D

cbielow commented 6 years ago

PTXQC 0.92.3 implements a reporter intensity metric, which covers labeling efficiency (indirectly, but sufficiently).