Compare three different conditions using 'scv.tl.rank_dynamical_genes'

[reneemoerkens]

Hello,

I have a scRNAseq dataset containing three different conditions and I want to compare the differentiation trajectories of these conditions. I have read here that you recommend running the velocities for the different conditions separately and then merging the datasets (I have used the concatenate function). Then I can use 'scv.tl.rank_velocity_genes' or 'scv.tl.rank_dynamical_genes' to compare the overall conditions. However, after merging, the 'scv.tl.rank_dynamical_genes' gives an error related to 'KeyError=fit_alpha'. Is this the correct method I am applying and how could I resolve the error?

Also, would it make sense to run latent.time on the merged dataset to get the same same root and end states in each condition (which shouldn't differ in my case between conditions)? And then compare genes that drive this trajectory in the different conditions?

Thank you!