Batch effet

[RunyuXia]

Hi! I’ve been running pySCENIC on my integrated mouse cell atlas using raw counts, but when I look at the UMAP generated from the AUC scores, I’m seeing strong batch effects across the dataset. Would it make sense to run pySCENIC separately for each batch and then assess the overlap in regulons, or is there a better approach to mitigate these batch effects? Thank you so much for your help!