I have a question regarding re-clustering after removing cells using SubsetData (R - Seurat package).

I am in the process of analyzing a relatively large single-cell dataset (16 separate samples of ~5-10k cells each). In our first run of the Seurat pipeline, we ran multiCCA to align/generate CCs from the 16 datasets. Subsequent analysis was performed using the ‘large Seruat’ output file generated from multiCCA.

Next, we wanted to eliminate unwanted cell types/clusters from this dataset. To do so, I used SubsetData to remove several clusters, creating a new ‘large Seurat’ file. Given that the dataset was generated using CCA, I can’t simply re-run FindVariableGenes and ScaleData, and then re-cluster (which would require running PCA)…

In our case, would the next steps be: 1) Identifying the individual cellIDs that were removed with SubsetData 2) Locating those cells in the original (pre-CCA) ‘large Seurat’ files 3) Deleting the unwanted cells there 4) Repeating FindVariableGenes and ScaleData 5) Re-running CCA across the manually filtered 16 samples

Is there a better/simpler way to go about doing this?

Thank you!