You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I was exploring the anndata objects created by scVelo and noticed that there is a key in obsm called "velocity_umap". This key doesn't seem to be used in any of the tutorials, but seems to be created during the scv.pl.velocity_embedding_stream(adata, basis='umap') step, the "umap" embedding looks as I would expect it to for the test data (computed with scanpy), but when plotting the "velocity_umap" it has a strange structure in both the datasets I looked at (the Dentate gyrus test data, and some outside data). Here are two examples.
Standard UMAP:
Velocity UMAP:
Standard UMAP:
Velocity UMAP:
It would appear that the cells with the largest velocity vectors are embedding around the outside of the velocity_umap plot, so I assume that this embedding has something to do with maybe the gene moments used to compute the velocity vectors rather than the gene expression?
Considering that there doesn't seem to be particularly clean clustering by cell type in the core of that velocity_umap graph and that they seem to have ~zero velocity, this being a UMAP based on some velocity measure rather than the expression would track with the key name. I'm largely wondering if this plot is something that people using the software need to be actively evaluating (if it might be informative for determining transition driver genes) or if it's best left alone. I didn't see any discussion of this embedding elsewhere so thought I'd ask about it.
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
-
Hello,
I was exploring the anndata objects created by scVelo and noticed that there is a key in obsm called "velocity_umap". This key doesn't seem to be used in any of the tutorials, but seems to be created during the
scv.pl.velocity_embedding_stream(adata, basis='umap')
step, the "umap" embedding looks as I would expect it to for the test data (computed with scanpy), but when plotting the "velocity_umap" it has a strange structure in both the datasets I looked at (the Dentate gyrus test data, and some outside data). Here are two examples.Standard UMAP:
Velocity UMAP:
Standard UMAP:
Velocity UMAP:
It would appear that the cells with the largest velocity vectors are embedding around the outside of the velocity_umap plot, so I assume that this embedding has something to do with maybe the gene moments used to compute the velocity vectors rather than the gene expression?
Considering that there doesn't seem to be particularly clean clustering by cell type in the core of that velocity_umap graph and that they seem to have ~zero velocity, this being a UMAP based on some velocity measure rather than the expression would track with the key name. I'm largely wondering if this plot is something that people using the software need to be actively evaluating (if it might be informative for determining transition driver genes) or if it's best left alone. I didn't see any discussion of this embedding elsewhere so thought I'd ask about it.
Beta Was this translation helpful? Give feedback.
All reactions