You can treat a SingleCellExperiment just like a standard R data.frame. The columns would allow subsetting on specific cells and the rows would allow subsetting on specific genes.
Here I subset my SingleCellExperiment object sce for only cells that contain the string WT.
sce <- sce[,grep('WT',colnames(sce))]
If you want to filter based on specific gene expression you could do something like this - keeping only cells with over 5k beta-actin reads: