Velocity/pseudotime inconsistencies

[MoLuLuMo]

Hi, I worked on PBMC dataset (T Lymph, 950 genes after filtering). Based on stream plot of velocity, I can see the arrows point from naive T cell toward memory & effector T cell. But pseudo-time plot shows reversed direction. Would that be possible to infer velocity pseudo-time (same as scvelo package)?

[danielee0707]

Hi. We don't usually recommend running RNA velocity methods for datasets such as PBMC because those consist of largely terminal non-differentiating cell types. MultiVelo doesn't natively support velocity pseudotime, but you can use scVelo's method directly by running scv.tl.velocity_pseudotime(adata_result, vkey='velo_s_norm') after running MultiVelo velocity graph function.

[MoLuLuMo]

Thanks. I only used naive, memory and effector CD4 T lymph cells for RNA velocity analysis. Theoretically, the trajectory should be clear