Re: [Psidev-ms-dev] mzQuantML update

[Jones, A. <And...@li...>]

Hi all,

Prior to the call today, I’m re-sending my mail that followed our last development workshop as a point for discussion. Some example files have since been updated in the googlecode repository, and there has been further discussion of scope since I realise we are not all 100% agreed on this yet, but this is gives us something to start the discussions.

Cheers
Andy



From: Jones, Andy [mailto:And...@li...]
Sent: 03 September 2010 11:12
To: 'psi...@li...'
Cc: 'psi...@li...'
Subject: [Psidev-ms-dev] mzQuantML update

Hi all,

Just a quick update for those who were unable to attend the recent mzQuantML meeting, Juan Antonio is going to circulate the full minutes later on. Hopefully we can arrange the next meeting at a time and place where others can attend, especially those from US, since we’re of course very keen to take on board input from all interested parties!

At the meeting we reached a tentative agreements on development approach:


-          We plan for development in stages, where we aim first for a mzQuantML level 1 that captures Protein and Peptide abundance values, but no models of underlying MS features or matches. MS features and feature matches will be added in Level 2. The rational is that the community needs a working data exchange model in the short term, most importantly to support database submission and data reporting.

-          Plan for monthly development conference calls starting from September, open to any interested parties. Date of first meeting to be set, but 4pm UK time on a Thursday seemed to be a good slot for meeting participants. If you would like to join these and would prefer another slot, let me know.

Most time at the meeting was spent working on data models for reporting protein and peptide relative/absolute abundance, grouping across replicates and protocols for specifying how these measures were generated. All artefacts have been (or will be soon) loaded into the SVN on googlecode: http://code.google.com/p/mzquantml/. We have uploaded a version 0.1 schema and some example files, although the repository and example files need tidying up over the coming weeks. Older example files are still archived there, so look at the dates files were uploaded to see which are recent.

It should be stressed that there are of course many open issues and nothing is seen as being close to a final modelling decision, so those who were not able to attend the meeting should get involved and push alternative ideas.

In summary, the model has the following structure at present:


ProteinList
     ProteinAmbiguityGroup

-          Optional quant values
 ProteinDetectionHypothesis

o   Optional quant values

o   peptideRefs

PeptideList
     Peptide

-          Optional quant values

AssayList (summarises collections of raw files)

-          QuantUnit (looking for a better name – encapsulates concept of one “run” on a label-free or one “channel” of a SILAC or iTRAQ, to be referenced elsewhere)

StudyVariablesList (groupings of QuantUnits – for purposes of grouping replicate values at PAG, PDH or PeptideLevel; this model has some recognised flaws that need to be fixed)

ProtocolList (definitions of how measures or ratios are calculated, for example with references to QuantUnits)

The optional quant values at PAG, PDH or PeptideLevel can be given for single QuantUnits or for StudyVariables (i.e. groups of replicates), and these can be single measures of abundance (plus associated stats) or ratios of abundance (plus associated stats).

Hopefully, this should spark some discussion on the list and then let’s aim for a first conference call towards the end of the month,
Best wishes
Andy