Filter to annotated cells

[mfansler]

In cases of reanalyzing published data, it may be preferable to filter to only annotated cells. This could be especially helpful in large datasets that may already exclude cells for poor QC metrics.

There's a TODO note here:

scUTRquant/scripts/mtxs_to_sce_txs.R

Lines 115 to 131 in e793dcb

	## Load MTXs Function
	load_mtx_to_sce <- function (mtxFile, bxFile, txFile, sample_id) {
	## TODO: Add annotation-based filtering
	## valid_cells <- filter(df_annots, sample_id == sample_id) %$% cell_id
	readMM(mtxFile) %>%
	as("CsparseMatrix") %>%
	`dimnames<-`(list(
	cell_id=str_c(sample_id, "_", read_lines(bxFile)),
	transcript_id=read_lines(txFile))) %>%
	t %>%
	{ .[, colSums(.) >= as.integer(snakemake@params$min_umis)] } %>%
	optional_hdf5_convert() %>%
	{ SingleCellExperiment(assays=list(counts=.),
	colData=DataFrame(cell_id=colnames(.),
	sample_id=sample_id,
	row.names=colnames(.))) }
	}