About simulation data

[mugpeng]

Hi,

I've learned a lot from your work about milo. But there are still some problems when I try to mimic your steps.

Here I am trying create some simulation data by Splatter, but I find it‘s really limited that can't control the proportion of condition labels in a specific part of one cluster or trajectory.

But, I would like to form the result like yours which a proportion of each branch or cluster between 10% and 80% of the data to be enriched or depleted by 25%.

Your method part mentioned:

The ground truth experimental signal was created using custom Python scripts, taking the ‘Steps’ latent variable from Splatter and randomly selecting a proportion of each branch or cluster between 10% and 80% of the data to be enriched or depleted by 25%.

So could you share your scripts or describe the main procedure?

Thanks a lot.

[uthsavc]

+1