search

HumanCellAtlas / secondary-analysis

Secondary Analysis Service of the Human Cell Atlas Data Coordination Platform
https://pipelines.data.humancellatlas.org/ui/
BSD 3-Clause "New" or "Revised" License
3 stars 2 forks source link

Analyze mouse 10x v2 dataset: HumanColonicMesenchymeIBD #816

Closed jychien closed 5 years ago

jychien commented 5 years ago

The "HumanColonicMesenchymeIBD" mouse bundles were previously run with optimus_v1.3.2 and need to be rerun with optimus_v1.3.5 because v1.3.3 was the one that was approved.

Project title: Structural Remodeling of the Human Colonic Mesenchyme in Inflammatory Bowel Disease Project short name: HumanColonicMesenchymeIBD Project UUID: f8aa201c-4ff1-45a4-890e-840d63459ca2 Metadata: mouse, 10x v2 Bundles: 6 primary bundles

https://humancellatlas.slack.com/archives/C6C5S6B5W/p1570230566137700

https://tracker.data.humancellatlas.org/?projectUUID=f8aa201c-4ff1-45a4-890e-840d63459ca2

@HumanCellAtlas/data-ops

samanehsan commented 5 years ago

@jychien since we do not have a way to update existing analysis bundles with new analysis results, we would have to wait until that feature is complete or tombstone the existing analysis bundles.

Suggestions on what to do here? @kbergin @jkaneria @jahilton

jychien commented 5 years ago

@samanehsan Thanks for the update. From a DataOps perspective, it would make more sense to tombstone the existing analysis bundles rather than to wait for feature development. The earlier we make the "new" analysis bundles available, we decrease the possible numbers of users that will be using the "old" analysis data.

kbergin commented 5 years ago

I can confirm tomorrow, but I don’t think any significant changes went in between 1.3.2 and 1.3.3 (hence only the patch version update). The only difference was a memory bump. 1.3.2 did include an important change from the previous release that was part of our validation. 1.3.3 just happened to be released when we got the mouse validation. I will work on us publishing a bit more information about our versioning process, ideally a patch version change should never impact whether something is valid scientifically.

As long as the mouse reference used in the analysis of this dataset with optimus 1.3.2 is the same as the one we are currently using in production then I would keep the 1.3.2 results and call this done.

kbergin commented 5 years ago

Confirming that there was only a memory bump between these two versions.

samanehsan commented 5 years ago

Thanks @kbergin! In that case we should keep the current analysis results instead of tombstoning them in production. @jychien is it okay to close this ticket?

jychien commented 5 years ago

Thanks @samanehsan and @kbergin! Yes, let's keep the current analysis and close the ticket.