Closed matthewspeir closed 2 years ago
The labels are currently drawn in the middle of all the cluster cells. So if these cells are not really clustered in the plot, it won't work or create weird labels. Do we have an actual example of a dataset where we need this? If not, I'd push it back for now as I haven't come across a dataset with two label fields yet.
Hmmm, can you maybe explain a little more how you would expect this feature to be used? I'm still having a hard time imagining it.
We had one dataset with two different sets of cluster labels, one was just 1-10 or so and the other field had the real cell type names. Jim thought it would be good to be able to flip between both. I’m not sure how I would put this into the user interface though ... and how common it is...
On Tue 2 Apr 2019 at 21:37, Matt Speir notifications@github.com wrote:
Hmmm, can you maybe explain a little more how you would expect this feature to be used? I'm still having a hard time imagining it.
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/maximilianh/cellBrowser/issues/89#issuecomment-479164805, or mute the thread https://github.com/notifications/unsubscribe-auth/AAS-TffJDazGwG9C84ksckvVPi3H5uBeks5vc7GCgaJpZM4cCB-2 .
Ahhh, that makes sense now. I agree I'm not sure how common this would be.
I have a single dataset with this. Have you ever seen it ?
On Tue 2 Apr 2019 at 23:16, Matt Speir notifications@github.com wrote:
Ahhh, that makes sense now. I agree I'm not sure how common this would be.
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/maximilianh/cellBrowser/issues/89#issuecomment-479206243, or mute the thread https://github.com/notifications/unsubscribe-auth/AAS-TeeVU4bGPr9tGs7ieDxR2hYmLWLpks5vc8iYgaJpZM4cCB-2 .
I don't think so. Which dataset is it?
The ciona one. Oops. I should push it...
On Tue, Apr 2, 2019 at 11:20 PM Matt Speir notifications@github.com wrote:
I don't think so. Which dataset is it?
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/maximilianh/cellBrowser/issues/89#issuecomment-479207934, or mute the thread https://github.com/notifications/unsubscribe-auth/AAS-TSIMf0h8Jw-c8bSbi8_Cp7cTanyPks5vc8mkgaJpZM4cCB-2 .
@maximilianh Not sure if this would fall into this feature request, but is it possible to define a cluster field to be used with a specific set of coordinates? e.g. the mouse gastrulation dataset has a few different clustering views, each one has a separate set of clustering labels/coloring that should be applied: https://hgwdev.gi.ucsc.edu/~mspeir/cb_demo/?ds=mouse_gastrulation
Yes, that's a configuration field, in the coords object, you can set "colorOnMeta":"fieldName".
example:
coords=[ {"file":"tsne.coords.tsv", "shortLabel":"t-SNE on WGCNA"},
changed to another one {"file":"subset.coords.tsv", "shortLabel":"neural cells", colorOnMeta="neuralCluster"}, ]
I added a paragraph about this now also to the "Advanced" docs page. One day, we'll re-arrange this page, but for now I'm simply collecting things that have no dedicated page yet.
On Wed, Apr 24, 2019 at 10:51 PM Matt Speir notifications@github.com wrote:
@maximilianh https://github.com/maximilianh Not sure if this would fall into this feature request, but is it possible to define a cluster field to be used with a specific set of coordinates? e.g. the mouse gastrulation dataset has a few different clustering views, each one has a separate set of clustering labels/coloring that should be applied: https://hgwdev.gi.ucsc.edu/~mspeir/cb_demo/?ds=mouse_gastrulation
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/maximilianh/cellBrowser/issues/89#issuecomment-486420678, or mute the thread https://github.com/notifications/unsubscribe-auth/AACL4TND5HX43MLZTPETK63PSDB3HANCNFSM4HAID63A .
I think the line you have in the example cellbrowser.conf is wrong. It has an equals sign instead of a colon.
Yes, I've fixed that a few days ago... oh shit, I haven't pushed it! Thanks!
On Fri, Apr 26, 2019 at 1:52 AM Matt Speir notifications@github.com wrote:
I think the line you have in the example cellbrowser.conf is wrong. It has an equals sign instead of a colon.
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/maximilianh/cellBrowser/issues/89#issuecomment-486877448, or mute the thread https://github.com/notifications/unsubscribe-auth/AACL4TPFOZEGCWKXGGFVE3TPSI73BANCNFSM4HAID63A .
So I did push now. But we still need some way in the UI to change the cluster label field. So not closing.
On Fri, Apr 26, 2019 at 2:25 PM Maximilian Haeussler maximilianh@gmail.com wrote:
Yes, I've fixed that a few days ago... oh shit, I haven't pushed it! Thanks!
On Fri, Apr 26, 2019 at 1:52 AM Matt Speir notifications@github.com wrote:
I think the line you have in the example cellbrowser.conf is wrong. It has an equals sign instead of a colon.
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/maximilianh/cellBrowser/issues/89#issuecomment-486877448, or mute the thread https://github.com/notifications/unsubscribe-auth/AACL4TPFOZEGCWKXGGFVE3TPSI73BANCNFSM4HAID63A .
This is all done now. There is a new UI now where you can select the label field. Can we close this, @matthewspeir ?
Yeah, I think we can close this now that we have the 'label by annotation' feature out there.
Define multiple cluster labels. @maximilianh should know what this means