Question on scv.pl.velocity_graph Lines

[Jennifer-Foltz]

Good afternoon!

Thank you for creating scvelo, we are very excited about using it! We have implemented the following code and are confused about the scv.pl.velocity_graph interpretation.

scv.pp.filter_and_normalize(sample, min_shared_counts = 20, n_top_genes = 2000)
scv.pp.moments(sample, n_pcs = 30, n_neighbors = 30)
scv.tl.velocity(sample, mode = 'stochastic')
scv.tl.velocity_graph(sample)

# scVelo plotting
scv.pl.scatter(sample, basis = 'X_umap', color = metadata_column, save = 'scatter.png', legend_loc = 'right margin')
scv.pl.velocity_embedding(sample, basis = 'X_umap', color = metadata_column, save = 'velocity_embedding.png', legend_loc = 'right margin')
scv.pl.velocity_embedding_grid(sample, basis = 'X_umap', color = metadata_column, save = 'velocity_embedding_grid.png', legend_loc = 'right margin')
scv.pl.velocity_embedding_stream(sample, basis = 'X_umap', color = metadata_column, save = 'velocity_embedding_stream.png', legend_loc = 'right margin')
scv.pl.velocity_graph(sample, basis = 'X_umap', color = metadata_column, save = 'velocity_graph.png', legend_loc = 'right margin')

After running, our UMAP has very few lines connecting cells in the velocity_graph, is this expected to occur or a bug?

Thank you!!

Jennifer