Sub Clustering an Integrated Cluster

[Abnaveed]

Hi Team,
I needed some advice on my data workflow that I do not seem to understand by reading past discussions about it.

I have 8 samples from 4 patients that were treated with a drug. We have pre-treatment and post-treatment single cell data for it. I followed the vignettes and get pretty good clusters after integration. My question is after I identify a cluster that I want sub-clusters for, what's the ideal work flow for it ? Do I subset the data and reintegrate it ? or Do I just rerun SCtransform and move to visualization and clustering without integrating it ? I read that re-integrating the data is not needed but I am pretty sure I am getting batch effects in sub-clusters if I just only rerun SCtransform and then find clusters.

Would appreciate any help or advice and thank you for this wonderful tool.