How to define root cells after diffusion map calculation

[MichaelPeibo]

Hi Adrian,
Thanks for providing such detailed code to reproduce your results and beautiful example of scRNA-seq analysis!

I am dealing with my own data. One issue came to me, I got two progenitor clusters(G1 phase and G2/M phase), very similar like your data set(some cluster and replicating cluster).

I wonder how to assign root cells for these two cluster?(merge two as one?)
I have been reading your code, however, it is a little complicated for me(very fresh in Python), could you explain that a little bit?

Thanks!

[MichaelPeibo]

an UPDATE:
I tried to filter genes highly correlated with TOP2A in highly variable genes like you described in the article. It mitigates the cell cycle effects when clustering.

Still, I wonder how you assign root cells.

Thanks!