Minutes_Standards_2021 09 - airr-community/airr-standards GitHub Wiki
- Are the changes in #541 sufficient to address the issues with Travis (#484, Call 2021-06)?
- Are there outstanding concerns with the single-cell representation (#409)?
- Any input for AIRR-C ExCom meeting on 2021-09-16 with WG updates?
- JSON vs. YAML: Continue discussion from Call 2021-08
- A growing number of tools are "AIRR compliant". Is this only support
for
Rearrangementor also forRepertoire? - Current status of the discussion on
Germline/Genotyperepresentation in the Schema (#530). - Review clonal abundance calculations (#543) and revisit
*_countfields (#161). - DataProcessing sprint is done, need to turn the discussion/results into action items and Github issues.
- Date: Mon, 2021-09-13 18:00 UTC
- Present: Brian, Chaim, Christian, Felix, Francisco, Jason, Katharina, Scott, Susanna
- Susanna will apply for Open Source Status for AIRR Standards repo for Docker. If this works, we could build a container for code testing there and pull into Github Actions. Jason will work on the latter. Alternatively we could apply for status as an Open Source project with Travis. If all else fails, see whether AIRR-C could pay the annual fees for these services (est. 1.2 kUSD p.a.).
- Single-cell representation (#409): Had a look at a "big" table TSV
(basically an n*m matrix with headers). Problem with headers as the
property names (e.g., "PAX5") are not known upfront. Also, some tools
might want to store multiple values (e.g., id, label, gid). Will
therefore try normalized approach instead
(
[cell_id, property, value]). Implementation for on-disk and the Schema for next call. Linkage on-disk is expected to happen via the Manifest file.cell_idshould only be considered unique in the context provided by the repository / created by the Manifest. This might become stricter in the future. - JSON/YAML: Agreement that tools MAY support YAML but SHALL support JSON. If users do not want to write JSON manually, they can write YAML and convert it (lossless). Need to update this in the Docs, which current suggest that YAML is an expected output format (#547).
- Should revisit "AIRR compliance" of tools: First, the list of applications supporting the rearrangement schema needs to be updated, as several of the tools still do not support the Rearrangement TSV. Second, the difference between "full" AIRR-compliance (as indicated by the badge awarded by the Software WG) and "just" supporting the Rearrangement TSV should be worked out, also in regards to support for other AIRR Schema objects. Will continue the discussion in the next call.
- Last three points of the agenda were pushed to the agenda of the next call (Call 2021-10.)
Copyright (c) 2017-2024 AIRR Community
