From 579a6144c78dbd54a3febe679ed7b1ff5ca3dd78 Mon Sep 17 00:00:00 2001 From: Mathias Walzer Date: Mon, 4 Sep 2017 11:53:03 +0100 Subject: [PATCH] Add telco August meeting notes I shall if time allows draft a short overview on encoding within PSI XML standards and the combination with element rule sets which stand in contrast with out of the box semantic validation and how our approach would compare --- meeting_notes/20170818_telco.md | 46 +++++++++++++++++++++++++++++++++ 1 file changed, 46 insertions(+) create mode 100644 meeting_notes/20170818_telco.md diff --git a/meeting_notes/20170818_telco.md b/meeting_notes/20170818_telco.md new file mode 100644 index 00000000..fb30e1d8 --- /dev/null +++ b/meeting_notes/20170818_telco.md @@ -0,0 +1,46 @@ +# QC working group teleconference 18 August 2017 + +## Attendants + +- Jinmeng Jia +- Douglas McCloskey +- Dave Tabb +- Stefan Tenzer +- Mathias Walzer + +--- + +# Coordination between the CV and MIAPE-QC +Picking up the conversation to coordinate the CV and MIAPE-QC as Jinmeng was unable to attend last month's teleconference. +* Dave asked what we learned in the last months on that combi of xml and json in regard to miape +* Mathias briefly recapitulated the use of json within value capturing elements, and that with the use of CV terms as 'keys' for value entries like the instrument params, MIAPE compliance checks are very feasible. +* Dave reminded that qcml is also ment to store analytical results (e.g. to detect which items are outliers) but not so much for replicating configuration files or mzid/mzq contents +* [Discussing the fusion of json formatting within XML might be a good agenda point for discussion (whether everybody feels this is a viable approach to implement) -- Ed.] + +# Shotgun minimum metric proposals +* Jinmeng presented an updated list of information items to be present in a minimum requirement for quality reporting. + - The list is much more more compact and a lot of detail went to the categorisation and direct connection to CV term groups for metrics. + - few issues were raised: + Dave gave a triple-tof data example -no percursor charge information- where it will be hard to compute a ratio; + Mathias gave for consideration the case if items are unavailable, e.g. report of missed cleavages if no enzyme used; + The agenda of no replication of search parameters (paramters from search settings should not be repeated in miape (e.g. missed cleavage) was emphasised by Dave; +* Jinmeng stressed the neccessity to reflect some form of experimental design, at least with groupings and/or replicates + - Dave points out the fact that it may be hard in certain situations to unambiguously sample measurements into fix categories of replicate + - Mathias notes that adds for consideration that providing a experimental design definition for qcML might be beyond the scope of a first version of qcML + - Dave suggests a simple hierarchy, level name/type arbitrary, which may be lab specific as exemplified with the CPTAC dataset. + - Mathias concurs that mapping predefined definitions to different levels of design may be achievable with CV terms + - Stefan argues that there should be a simple structure, to which every exp. design can be abstracted to + +# Discussion of the current XML schema +in addition to Wout's remarks in . +Mathias added remarks about RawFile element: +- raw meaning input file for spectral data; about naming, Dave argues that a majority is probably most familiar with calling it raw file) + + Missing a tag what kind of file it is, important when _not_ raw but mzML + + Checksum to indicate if qcml file reflects the data from the file at hand (filename may be inconclusive) +- + +# EuBIC developer's meetings project proposals +Agenda point untouched + +# Extra item +Douglas McCloskey joined the call for the first time and introduced the field of QC application for which he and Mathias have been discussing qcML adaption - isotope dilution mass spectrometry of metabolites.