RFC schedule and deployment plan

[Bankso]

To expand the MC2 data model, data type-specific metadata models should be assembled and reviewed by the community via the Request for Comments process. One strategy for implementing is as follows:

• identify priority metadata models (assay or data type, processing level) and key information that we want to be able to provide for the community (common pipelines/processing, common repos, common access requirements, etc. Basically, any question posed about data reuse/sharing should be considered as content to ask about during the RFC)
• assemble schedule of RFC dates
• hold an introductory webinar on the RFC process for the community, to discuss the guidelines, expected outcomes, rules of engagement, use case requirement, etc.

For each metadata model:

• select key community members to email about the RFC
• assemble version 0, based on existing models/standards (HTAN, NF, AD, CRDC, etc.)
• have a sheet with prelim model and RFC document available for contributors
• have the storage repo included (type-specific if possible)
• record common processing methods/pipelines
• for QC metrics, indicate the pipeline or tool used and the source doc or output

Some priority data types:

• biospecimen
• tools
• bulk RNA-seq level 1 - 4
• scRNA-seq level 1 - 4
• 10X Visium (or general spatial, if we can decide on how to put that together)
• imaging/microscopy (recommend stitched images and then masks and analysis, avoiding the giant raw images where possible, unless required for secondary use)

[Bankso]

Work here will be informed by #74, in particular, 1) the level of CDS/DataHub model integration we reach and 2) our perceived sufficiency of the CDS/DataHub models for MC2 use cases. If we primarily adopt the CDS/DataHub assay models, I'm not sure how critical RFCs will be

[aclayton555]

Expect this work to be within the scope of work for #115 and will be a specific output

[aclayton555]

Initial brainstorm with Milen, Ashley, and Orion: https://docs.google.com/document/d/1dF1-FjGSdO3nkKArEsrnjnWFLeOV78MlvGZvM8smJVk/edit

[aclayton555]

mid-sprint:

• start to work on design doc
• schedule for RFC to be informed by what we are receiving (e.g. imaging data likely first)
• need to start thinking about how we organize the data model (e.g. data types, processing levels)

[Bankso]

Work completed:

RFC tool design doc

RFC tool mockups

Data model visualizer for MC2

RFC tool app repo

Next steps:

• refine app UI
• finalize functions required to handle form data + github integration
• build API
• build github actions to handle API calls

[aclayton555]

24-7/8 close-out:

• @Bankso will do a bit of tidy up and share app prototype with Milen to assess whether support is available to improve data model visualization tool, and what the timelines would be. If immediate support is not available, could link users to data model docs for more information. This is what Aditi has done in HTAN.
• Ensure we build in a robust method for pulling in the latest version of the data model (should be able to do this similar to how the docs site works).
• thinks about (external) user experience and instructions for users
• incorporate "Attribute Required?" true/false

Roll this into the next sprint to complete next steps. A target will be to use this in a pilot with the Tools schema. Alts could be biospecimen. Timeline for October.

[aclayton555]

24-9: secondary to site visit priorities