coolneng
/
locigenesis
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Save v_call and j_call to a text file

This commit is contained in:
coolneng 2021-03-11 21:03:16 +01:00
parent 7b15df7614
commit 762791829a
Signed by: coolneng
GPG Key ID: 9893DA236405AF57
1 changed files with 23 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

35
src/repertoire.r
View File

@ -10,18 +10,30 @@ generate_repertoire <- function(number_of_sequences) {
))
}
# TODO save also v_call and j_call
preprocess_data <- function(repertoire, sequencing_runs) {
sequences <- as.character(repertoire$sequence)
reads <- Biostrings::DNAStringSet(rep(sequences, sequencing_runs))
names(reads) <- seq_len(length(reads))
reverse_complement <- Biostrings::reverseComplement(reads)
return(reverse_complement)
amplify_rows <- function(data, column, factor) {
if (column == "sequence") {
dna_string <- Biostrings::DNAStringSet(data)
reverse_complement <- Biostrings::reverseComplement(dna_string)
return(Biostrings::DNAStringSet(rep(reverse_complement, factor)))
}
return(as.character(rep(data, factor)))
}
save_data <- function(repertoire) {
file_name <- "data/sequence.fasta"
Biostrings::writeXStringSet(repertoire, file_name, format = "fasta")
save_data <- function(data, name) {
if (name == "sequence") {
file_name <- paste("data/", name, ".fasta", sep = "")
Biostrings::writeXStringSet(data, file_name, format = "fasta")
} else {
file_name <- paste("data/", name, ".txt", sep = "")
cat(data, file = file_name, sep = "\n")
}
}
process_data <- function(repertoire, sequencing_runs) {
columns <- c("sequence", "v_call", "j_call")
data <- repertoire[, columns]
amplified_data <- mapply(data, names(data), sequencing_runs, FUN = amplify_rows)
invisible(mapply(amplified_data, names(amplified_data), FUN = save_data))
}
parse_cli_arguments <- function(args) {
@ -36,5 +48,4 @@ arguments <- parse_cli_arguments(args)
number_of_sequences <- as.integer(arguments[1])
sequencing_runs <- as.integer(arguments[2])
repertoire <- generate_repertoire(number_of_sequences)
processed_data <- preprocess_data(repertoire, sequencing_runs)
save_data(processed_data)
process_data(repertoire, sequencing_runs)