ouija assumes small set of marker genes as prior information

[rcannood]

ouija assumes a small panel of marker genes chosen a priori! marker genes should also be a possible form of prior information

[zouter]

True, although we are going to have a look how to do this for real data (maybe predict the group of a cell based on its expression using a random forest and then selecting the most important genes?)

[zouter]

This will be added for the synthetic data as well :man_dancing:

[rcannood]

For real data, this could indeed be an option. The gold standard grouping could even be used.

[zouter]

Done