Scvelo vs Monocle3

Question

[Chris]

Hi Biostars,

I tried to do trajectory analysis but got an error message here Help with error velocyto so I tried Monocle3 and I feel it was a lot faster and simpler than using scVelo/velocyto. However, with Monocle3, how can we know where the root note to choose? Can I use trajectory analysis on data from phenomena such as Endothelial-Mesenchymal Transition (EndoMT)? Thank you so much!

[ATpoint]

Monocle does trajectory, scVelo does velocity. That's two different concepts. There is literature to explain the difference. In a nutshell, trajectory is (simplified) trying to find the shortest path from a root to end end across a number of datapoints. Velocity uses ratios of (un)spliced counts. What can be done with a method, as said in many posts before, all depends on your data, project, biological question, conditions, and people here cannot guide you through your project. So please try to ask technical questions agnostic of your project, because that is what people can help with.

To answer the technical part: slingshot (Bioconductor) can accept a a priori known start and end point and fit the trajectory accordingly. Maybe monocle can do, idk as I only used slingshot. Often the root would be known in my experience based on other analysis, such as markers for maturation states or something like this, so you can tell the tool where to start and where to end, rather than leaving it best-guessing.

[Chris]

Thanks ATpoint! Will try slingshot. I used FindAllMarker() to annotate each cluster but the cell types were not as I expected. Not cluster 1 is cell type A, cluster 2 is cell type B and we know cell type A will develop/transition into cell type B.