SED-ML Editors Meeting, 13 Oct 2017
Milan, Italy, Fri, 13. Oct 2017 at 15:00 CEST (COMBINE 2017)
Attendees: Matthias König, David Nickerson, Lucian Smith, Dagmar Waltemath
Topics
-
Summary of progress at COMBINE 2017: The editors reported on the progress made at the COMBINE
meeting, summarising the issues and discussion points which are also listed in the meeting notes
from the COMBINE 2017 Breakout Sessions.
Decision: The editors will update the L1V3 specification and send
an errata to sed-ml-discuss
(Clarification of the meaning of NA (not available) values; All elements should be derived from
SED-Base; Clarification of the meaning of the attributes numberOfPoints, therefore creation of a new
attribute numberOfSteps and deprecation of the attribute numberOfPoints)
The editors further agreed on including a number of suggestions for the L1V4.
-
The editors further agreed on including a number of suggestions for the L1V4 specifications, as
discussed at the COMBINE meeting:
-
Make @id optional on SED-Base and elements that require @id specify that in their
UML/schema.
Vote: YES
-
The editors postponed other decisions, when they saw need for further discussion in the wider
community, or for further implementation support for testing:
-
Need to be able to iterate over data sources, but need further testing of the proposal via
implementation to make sure it works (this would be in L1V4) - postponed
-
Extend Slice in DataDescription to include a start and end index to enable subsampling data -
postponed
-
Need a way to address data in “complex resources” - discussion whether the suggestion of adding
a Resource class is ready to vote or not, move toward using URIs to address into archives, etc.
- Decision: involve sed-ml-discuss
-
Specifying output file names - The editors remembered from previous meetings that this is not a
popular addition, rather out of scope for SED-ML, but topic for a workflow definition. @id on
Outputs could be used to define an output location. Decision: This suggestion is currently not
ready for addition to L1V4
Decision: should be brought up on sed-ml-discuss by the
interested parties.
-
DimensionDescription should be extended to support various formats (currently CSV and NuML)
Vote: Add identifiers for excel and hdf5 and put on the list
on SED-ML website: YES
-
Design plotting styles that are defined and can be “reused” (line styles, colours, symbols, etc)
- This was discussed among the editors, but no clear task to vote on this.
-
Revision of output definitions - This was discussed among the editors, but needs to be discussed
further with the community and a specific proposal needs to be designed & tried out in
implementations
-
Deprecate log-attribute on curves: the result of the breakout sessions at COMBINE was that a new
scale attribute should serve as an enumeration of possible values (log, log10, log2…)
-
Report on the discussion on the addition of a “figure” output type, allows subplots, etc.
Conclusion:
This needs further discussion on sed-ml-discuss.
-
Agreement amongst editors and during COMBINE breakout that NonUniformTimeCourse and
Interpolation are needed
Conclusion: Proposal needs fleshing out and discussing on
sed-ml-discuss.
-
At COMBINE, SED-ML Logical models have been discussed. Lucian Smith offered to add example models to
the existing test suite
-
The editors mention that many more issues have been raised on github (https://github.com/SED-ML/sed-ml/issues),
but these need to go through sed-ml-discuss before the editors can make decisions.
-
-
The editors talked about the existing, sparse list of SED-ML presentations on sed-ml.org, which
is not (yet) a helpful resource for people interested in SED-ML
Decision: Extend list at least with all presentations that
the editors are aware of.
-
The editors further discussed about adding the dates of latest releases on the page listing SED-ML
specifications
Decision: Dates for stable releases yes, but not dates for
drafts.
-
The editors decided to promote the release of L1V3 (during COMBINE) on
twitter through the COMBINE Coordination account.
-
The editors furthermore plan to provide a longer news item for sed-ml.org
-
The editors discussed about publishing L1V3 in the special issue on standards that appears
annually in the Journal of Integrative Bioinformatics - Decision: The editors aim for publishing
L1V3 in that journal and will prepare the manuscript as soon as the call is out, presumably be
the end of 2018.
-
Matthias König reported on plans of extending the model of the
month provided by BioModels. He would prepare COMBINE archives (including SED-ML files)
whenever possible and provide links to tools that can process the archives
Conclusion: The editors agree that this would be beneficial
for the community.
-
Further topics
-
Publication of editor’s meeting notes: The editors discussed publishing their meeting notes on
sed-ml.org, following the example of the SBGN Editors group - Decision: From now onwards,
meeting notes will be published on the About page.
-
Development of L2: The editors came to the conclusion that L1 gets more and more complex, and
some of the older classes from L1V1/2 may not be needed anymore - Conclusion: The editors
propose to stick to L1 for now to gain some more experience with the extensions currently being
discussed. The plan is then to design a new structure in L2, that may hopefully be more simple…
together with the community, using one of the following COMBINE meetings as a kick-off.
-
Editors election: Two editors will be substituted by the end of the year, and elections need to
be organised. The editors leaving the board this year will be responsible to organise the
election.
-
COMBINE Coord: The editors discussed who should represent the SED-ML editorial board in the
COMBINE Coordination board, and several editors volunteered - Decision: The editors will consult
the COMBINE Coordination Board to seek advice on how to proceed.
The next editor’s meeting is scheduled for November 2017.