Add SKC documentation page - #509
Conversation
sjspielman
left a comment
sjspielman
left a comment
There was a problem hiding this comment.
This looks fine to me and all decisions make sense. Left a few small comments but nothing major so I don't feel that I need to see this again. Also a couple thoughts thrown in there for good measure.
| @@ -0,0 +1,30 @@ | |||
| # The Sean Karl cohort | |||
|
|
|||
| This page contains information about additional processing specific to the Sean Karl cohort, which corresponds to samples available as part of `SCPCP000026`. | |||
There was a problem hiding this comment.
confirming this is correct; your PR comment said 27, but it's indeed 26 :)
| To generate these custom cell type annotations, tumor cells were identified using `ews-nf` with the default annotation workflow (`--workflow annotation`). | ||
| Please see the [`ews-nf` documentation](https://github.com/AlexsLemonade/ews-nf/blob/main/README.md) for more information on how tumor cells were classified. | ||
|
|
||
| All immune cells (identified by their {ref}`consensus cell type annotations <processing_information:Cell type annotation>`) were then further refined by the submitters. |
There was a problem hiding this comment.
Eventually we might have a citation here I suppose
|
|
||
| ## Metaprograms | ||
|
|
||
| Recurrent gene expression programs, or metaprograms, were identified across all tumor cells and all samples within the Sean Karl cohort using the metaprogram workflow within `ews-nf` (`--workflow nmf_metaprograms`). |
There was a problem hiding this comment.
General question - is this going to stay in ews-nf? Or will it eventually be fully removed and only available standalone? Assuming the former?
There was a problem hiding this comment.
General question - is this going to stay in ews-nf? Or will it eventually be fully removed and only available standalone? Assuming the former?
This is going to stay in ews-nf as is.
|
|
||
| Recurrent gene expression programs, or metaprograms, were identified across all tumor cells and all samples within the Sean Karl cohort using the metaprogram workflow within `ews-nf` (`--workflow nmf_metaprograms`). | ||
| Briefly, `cNMF` ([Kotlier _et al._ (2019)](https://doi.org/10.7554/eLife.43803)) was used to run non-negative matrix factorization (NMF) on each sample across a range of ranks. | ||
| The resulting NMF programs across all samples were then clustered into metaprograms based on their Pearson correlation coefficient. |
There was a problem hiding this comment.
No mention spectra?
Just kidding, do not mention spectra.
| <!--TODO: Update the naming scheme here with anything decided in https://github.com/AlexsLemonade/scpca-nf/pull/1307--> | ||
| Column names are formatted with: `submitter_data_<metaprogram name>_score`, where the `metaprogram name` is the designated name provided by the submitters. | ||
|
|
||
| For more detailed information on how metaprograms were generated, see the [`ews-nf` documentation](https://github.com/AlexsLemonade/ews-nf/blob/main/README.md) |
There was a problem hiding this comment.
Something to consider to circle back to - do we want to include a tagged version of this repo in these docs? Not sure.
There was a problem hiding this comment.
Hmm... I want to say yes, but if we ever make changes to that workflow and regenerate results to add to the Portal then we would have to remember to update the link here which seems hard to maintain.
Co-authored-by: Stephanie J. Spielman <stephanie.spielman@gmail.com>
|
Just noting that I updated this in 939377d to add a sentence with a link to the main SCE contents page to provide context that the included information is in addition to all standard contents. |
2 participants
Closes #506
This PR adds a new page that is just for the samples that are part of the SKC (SCPCP000027). The page itself is hidden so that it's not in the table of contents or navigation bar, but it is still searchable. This means users can still search for the SKC and pull up the page, which I think is okay. The link to this page will be included on the project page under the "Additional processing" section and in the readme for this specific project.
As for the contents of the page, I included a section on the custom cell types and then a section on the metaprograms. I didn't go into too much detail behind the methods since a lot of that will be found in the
ews-nfdocumentation, but enough to at least tell users where the information is coming from. I mostly focused on making sure they knew where in the object information was stored.We also have an open PR with some discussion around the exact name of these columns, but I think we have landed on
submitter_data_for the prefix (see AlexsLemonade/scpca-nf#1307 (comment)). I'm going to give instructions to Tony for transfer to ensure that the columns that he sends us have<metaprogram name>_scorewhere the name is the biological name that they are working on assigning. Let me know if you disagree with this plan? I included a TODO just in case the name format changes.Also, I included links to
ews-nfthat are currently private but will be public before this goes on the Portal.