R/iseqr_make_tcr.R

In ahopki14/tcrSeqR: Tools for analyzing data from Adaptive's ImmunoSeq platform

#' iseqr_make_tcr
#' 
#' Make a tcr object from an imported adaptive dataset and a dictionary
#' containing sample metadata.
#'
#' @param ds A dataset containing Adaptive TCR data, made with `iseqr_merge`
#' @param dict A dictionary containing the sample metadata. This must contain a
#' field called `fn` with sample names matching the original tsv files (or, more
#' accurately the colnames(ds)).
#' @param remove Logical, indicating if non-productive clones and clones with
#' stop codons should be removed
#' 
#' @return A `tcr` object 
#' @author Alexander Hopkins
#' @export

iseqr_make_tcr <- function(ds, dict, remove=T){
	stopifnot(class(ds)=='data.frame')
	# restrict to productive and sanitize factor
	if(remove){
		w_remove <- unique(c(which(ds$aa==''), grep('\\*',ds$aa)))
		ds <- ds[-w_remove,]
		cat('Removing',length(w_remove), 'clones with stop codons or no translation\n')
		ds$aa <- as.factor(as.character(ds$aa))
	}
	#reoder the dictionary to match the ds
	dict <- iseqr_order(dict, ds, reorder=T)
	#make an ordering variable
	dict$ord <- seq(nrow(dict))
	#get metadata and data column locations
	w_m <- grep('aa|nt|syn', names(ds))
	w_d <- grep('aa|nt|syn', names(ds), invert=T)
	# make the SE object with the data
	if(!any(grepl('syn',names(ds)))){
		#if nt data
		ds_se <- SummarizedExperiment(assays=list(tcr_nt=as.matrix(ds[,w_d], rownames=F)))
				}else{
		#if agg data
		ds_se <- SummarizedExperiment(assays=list(tcr=as.matrix(ds[,w_d], rownames=F)))
	}
	# add some metadata
	rownames(ds_se) <- ds$aa
	rowData(ds_se) <- DataFrame(ds[,w_m])
	stopifnot(all(colnames(ds_se)==dict$fn))
	colData(ds_se) <- DataFrame(dict)
	colnames(ds_se) <- colnames(ds[,w_d])
	#return the SE object
	tcr(ds_se)
}