After sample integration there are often cells from one sample that and inappropriately clustered with cells from another sample. In my experience this has generally been when a cell type is exclusive to one sample.
The best way that I have found to resolve this is to identify it and then redo the integration excluding the exclusive cluster. Clusters can currently be excluded during integration. This would permit identifying when the issue has occurred.
After sample integration there are often cells from one sample that and inappropriately clustered with cells from another sample. In my experience this has generally been when a cell type is exclusive to one sample.
The best way that I have found to resolve this is to identify it and then redo the integration excluding the exclusive cluster. Clusters can currently be excluded during integration. This would permit identifying when the issue has occurred.