Removing contamination of set of genes from selected cell clusters

Question

[singcell]

I am comparing WT and KO samples using scRNA-seq integration vignette. As expected, there is a cell cluster which has lower number of cells in KO but very high numbers in WT. However, expression of the genes specific to the cluster in WT is very high and its expression shows up in all WT clusters. Which I believe is a contamination of the transcript from high expressing cells in all other cell types. Is it possible to remove this contaminating transcript/transcripts from other cell cluster (except the one where it belongs) without affecting expression data? What would be the vignette to use?

[ATpoint]

Words alone are not suitable to describe your problem. Show plots and code. Remember, we cannot see your screen, nor know your code, analysis, experimental setup.