search

AlexsLemonade / scpca-docs

User information about ScPCA processing
https://scpca.readthedocs.io/en/latest/
BSD 3-Clause "New" or "Revised" License
0 stars 1 forks source link

Update metadata file description for multiplexed data #367

Closed sjspielman closed 1 week ago

sjspielman commented 1 week ago

We recently added a column demux_cell_count_estimate to the multiplexed metadata. Unfortunately, this made it into the docs as demux_sample_cell_estimate:

This issue tracks correcting this column name in the docs.

jaclyn-taroni commented 1 week ago

It's not the docs that are wrong: https://data-lab-knowledge.slab.com/posts/2024-10-01-sample-count-discussion-5o843itw

sjspielman commented 1 week ago

It's not the docs that are wrong:

Well, now I don't understand this discrepancy. Is it even a discrepancy? Context: https://github.com/AlexsLemonade/OpenScPCA-analysis/pull/880#discussion_r1842759671

sjspielman commented 1 week ago

Ah, is the implication scpca-nf is recording this wrong then?

jaclyn-taroni commented 1 week ago

I think the download is not what was discussed as recorded in that Slab document (or at least as I understood it).

jashapiro commented 1 week ago

Yes, it appears that the field name in slab was not used in the portal code. At this point, it seems the easiest change to make is the docs, especially given that the current field name isn't particularly bad.

jaclyn-taroni commented 1 week ago

The other question would be if the calculation is correct. If it is correct, and it's just the name, we can just update the docs.

jashapiro commented 1 week ago

The other question would be if the calculation is correct. If it is correct, and it's just the name, we can just update the docs.

The calculation does appear to be correct. At the moment, there are some small (1-2 cell) discrepancies between what is on the portal and what was is in the input bucket, but I expect these are due to the changes associated with rounding. (The portal does not seem to yet have the updated SCPCP000009 files, confirmed by downloading a sample and seeing non-integer values)

jaclyn-taroni commented 1 week ago

I propose we do the following:

This doesn't impact downloadable files – we are merely correcting documentation to reflect the current downloadable files – so I'm not sure a CHANGELOG entry makes sense. Does anyone disagree?