Run cicero on multiple merged cds objects

Daniel Gingerich

unread,

Oct 12, 2021, 4:44:55 PM10/12/21

to cicero-users

Hello,

I was originally running cicero on 24 different datasets - 12 control, 12 disease. I want to perform a differential network analysis like in WGCNA, to identify differentially active cis reg elements. For this, I would need to preserve the metadata information. When the cells are aggregated, metadata is no longer linked to the data. Is it valid to create separate cicero cds's for disease and control, then merge the two and run the algorithm?

Best,

Dan

Daniel Gingerich

unread,

Oct 12, 2021, 4:48:13 PM10/12/21

to cicero-users

If this is ok to do, I think an enhancement for a future cicero release would be to add a cell binning option that keeps specified variables binned separately. This would aid in future analysis of cicero CCANs

hpl...@gmail.com

unread,

Oct 14, 2021, 1:56:48 PM10/14/21

to cicero-users

Hi Dan,

Yes I don't see why this wouldn't be valid, just be careful that you're using the same peak sets or you won't be able to merge afterwards. combine_cds from monocle3 should be able to handle the merge. That's a good idea for an enhancement, I'll add it to the issue list for the next time I do a bigger update.

Best,

Hannah

Message has been deleted

Daniel Gingerich

unread,

Oct 15, 2021, 9:59:11 AM10/15/21

to cicero-users

Hannah,

Thanks for the input! I appreciate that you are considering adding that to the next update!