Re: [Psidev-qc-dev] HUPO-PSI Quality Control Working Group

[Robert C. <cha...@cg...>]

Hi David,

The publication guidelines are for making authors supply minimal 
information to allow assessment of the reliability of results. There is 
not necessarily a direct relationship between the quality of the data 
and quality of results: e.g. someone could have terrible chromatography 
and dirty samples, but still reliably identify all of the binding 
partners in an IP experiment.  Hence, for many studies the qcML file 
will not inform about the reliability of the results.  Where it does 
come into play is in comparative and quantitative studies; especially 
label-free studies, where chromatography and reproducible instrument 
performance are prerequisites.  I could see it as being required in SRM 
or other DIA studies, and could be added to our clinical guidelines.
What is not clear to me is whether this is something that the authors 
need to produce, or whether the repositories (PRIDE; MassIVE; PASSEL) 
should automatically create them.  There is also the ‘lowest common 
denominator’ problem of only being able to ask for it can be produced 
from all instrument files.  Is this derived from mzML files?
MCP has an Associate Editors meeting at the end of January.  I could 
potentially bring this up as a topic of discussion during the day if the 
paper is published by then, or if you send me it in whatever state it is 
in at the beginning of the New Year.

Robert


On 11/30/2016 4:19 AM, Tabb, David, Prof <dt...@su...> wrote:
>
> Hi, Sue and Robert.
>
> I am writing in connection with my work in the HUPO-PSI Quality 
> Control Working Group.  We were founded in April as a way to 
> facilitate the use of quality control workflows with proteomics and 
> metabolomics data sets.  Several groups have now produced tools to 
> generate proteomics quality metrics from raw data or raw data plus 
> identifications.  Our working group hopes to put forward standard ways 
> to communicate these metrics (particularly a file in an updated “qcML” 
> format), making it possible to make downstream decision-support tools 
> that are able to work with a variety of front-end metric generators.
>
> We believe one of the important ways we can leverage these metrics is 
> by ensuring that data sets that reach publication have some basic 
> quality information accompanying them (for some tools, this would mean 
> spending just a few minutes of CPU time to generate the qcML 
> document).  The qcML document could, for example, inform potential 
> users of the numbers of MS/MS scans acquired, the proportion of 
> precursors that were +2, or the distribution of TIC throughout the MS 
> scans.
>
> We submitted a manuscript to /Analytical Chemistry/ that introduced 
> the goals of the working group to solicit input from the broader 
> community.  A reviewer has suggested that we need to show greater 
> clarity on how we will interact with journals to nudge publication 
> guidelines to recommend the inclusion of qcML once the format is 
> finalized and the tools to produce them are easily acquired.  The two 
> of you came to my mind as great resources on how this evolution could 
> take place.
>
> Do you have some insights on how receptive ACS and ASBMB journals 
> would be to including this recommendation for proteomics and 
> metabolomics manuscripts?
>
> Thank you,
>
> Dave
>

-- 
Robert Chalkley PhD
Adjunct Professor
Genentech Hall, N474A
Tel: 415 476 5189
Fax: 415 502 1655

Mass Spectrometry Facility
University of California San Francisco
600 16th Street,
Genentech Hall, suite N472A
San Francisco, CA 94143-2240