PXD000279

[ypriverol]

Hi @daichengxin:

This dataset (issues #1 ) has one fraction missing RAW file breaking the quantms pipeline. In the current version of quantms this can't be run.

[ypriverol]

Results of the dataset PXD000279 can be seen here: http://ftp.pride.ebi.ac.uk/pride/data/proteomes/proteogenomics/benchmakrs/

The current results are worse than the original quant results from MQ. Some issues found it:

• [ ] L peptides in the database are not identified. Instead, the MQ results reports a lot of hits with I as a replacement.
• [ ] Lost of hits in the MQ results are not even identified in the non-filtered list from MS-GF+ and Comet. Example: Peptide: TSEFQAEVQK. Interestingly the scan number for this peptide in the MQ output is: 864524 However in the mzML converted with ThermoRawFileParser the highest number for one scan is 16168.

I have reviewed in details the parameters, the database used and all seems to be the same that the original MQ searches.

Here the peptides from MQ.

modificationSpecificPeptides-results.txt.gz

[jpfeuffer]

As long as the L peptides map to proteins that have both I and L, everything is fine. Of course, this needs to be considered during comparison, but only when comparing peptides.

864524 seems to be unlikely. Probably MQ is counting all spectra in all files. That is a pity for comparing. The only thing we could do is compare it to an MSconvert conversion. But such a big error would have been found by now I guess.

Can you describe "worse"?

[ypriverol]

We have decided to re-run all the searches again with MQ and see if the results are different. We have seen two completely difderent peptide spaces between the orignal searches in the paper and our results and this is strange. We will use same database and perform the searches again.