From 8f5b9ee698f688201de7bb197f0172ee30d9f970 Mon Sep 17 00:00:00 2001
From: coolneng <akasroua@gmail.com>
Date: Tue, 23 Mar 2021 20:54:31 +0100
Subject: [PATCH] Parse curesim and VDJ sequences from files
---
src/alignment.r | 17 +++++++++++++++++
1 file changed, 17 insertions(+)
create mode 100644 src/alignment.r
diff --git a/src/alignment.r b/src/alignment.r
new file mode 100644
index 0000000..ea6d966
--- /dev/null
+++ b/src/alignment.r
@@ -0,0 +1,17 @@
+library(Biostrings)
+library(fastqcr)
+
+parse_data <- function(files) {
+ reversed_sequences <- Biostrings::readQualityScaledDNAStringSet(files[1])
+ sequences <- Biostrings::reverseComplement(reversed_sequences)
+ vdj_alignment <- read.csv(files[2])
+ vdj_string_set <- lapply(vdj_alignment, FUN = Biostrings::DNAStringSet)
+ return(list(sequences, vdj_string_set))
+}
+
+align_sequences <- function(sequences, vdj_segments) {
+
+}
+
+input_files <- c("data/curesim_sequence.fastq", "data/vdj_alignment.csv")
+data <- parse_data(input_files)
\ No newline at end of file