A question regarding augmenting HVG with marcopolo

[revolvefire]

First of all, thank you for the wonderful tool.
I have been working on multiple samples from different developmental time points and was hoping that MarcoPolo could cluster out some of the developing cells with more accuracy.

I am not super familial with the scanpy pipeline and relatively new to the field so please kindly bear with my rudimentary questions.

I was wondering what specific steps are needed to augment HVG from the standard pipeline.
Am I supposed to simply replace .var.highly_variable from the scanpy pipeline with marker_results from MarcoPolo?
Also, if I am working with multiple samples, can I still adopt this process (augmented HVG) for the integrated data set (concatenated anndata)?

Thank you

[chanwkimlab]

Hi, sorry for the late reply. I missed the message. Yes, you are right. You can replace genes indicated in var.highly_variable with the genes with high MarcoPolo scores. Alternatively, you can even mix those two sets of genes as well. Please let me know if you have more questions.

Best,
Chanwoo