Problem with arrow direction [Question]

[pabloatria18]

Hello. I am sorry I know there are some other issues open that are similar to mine but I have not been able to fix this. I am totally new to RNA velocity, this being my first analyzed dataset.
I have a dataset from bone, where the most differentiated group of cells (Cluster 6) and the intermediate state (Cluster 2) is pointing towards the progenitors (Cluster 0-1-3).
I followed the directions that I found in other issues such as #112 and #173

Here I attach my outputs:
scv.pl.proportions(adata,groupby='Clusters')

UMAP

scv.tl.velocity(adata, mode='stochastic')

scv.tl.velocity(adata, mode='steady_state')

scv.tl.recover_dynamics(adata)
scv.tl.velocity(adata, mode='dynamical')

print(adata.var['velocity_genes'].sum(), adata.n_vars)
379 2000

top_genes = adata.var.sort_values('fit_likelihood', ascending=False).index[:10]

Color Pseudotime

counts_s = scv.utils.sum_var(adata.layers['spliced'])
counts_u = scv.utils.sum_var(adata.layers['unspliced'])
fractions_u = counts_u / (counts_s + counts_u)
scv.pl.scatter(adata, color=fractions_u, smooth=True)

I appreciate the help
Thank you very much!

[WeilerP]

@pabloatria18, please see #462 or #216. Happy to have a discussion there if something is still unclear.

[pabloatria18]

Thank you @WeilerP, Sorry for duplicating the issue!
I saw both issues you mentioned, and I noticed not all of my genes have the almond/football shape as mentioned.
I have 379 velocity genes and I was wondering if there is a way of selecting some of them and evaluate if the direction changes?
What would you recommend me to do in this case?
I really appreciate your time and advice!

[WeilerP]

@pabloatria18, can we please move the discussion to #462 to have the related problems in the same place? Just repost your question there. Thanks!

[pabloatria18]

Sure! Thank you @WeilerP!