- [Outside of pipeline] Demultiplex with bcl2fastq. After demultiplexing, there will be a pair of FASTQ files for each unique i5/i7 index (typically 96).
- Extract cell barcode and UMI from read 1, add to read 2 FASTQ. Validate barcodes against list of RT and ligation barcodes used. By default allows for 1 mismatch (I don't think this is actually implemented, although the variable is set to 1). Output in
UMI_attachfolder. - Trim A+ sequence from FASTQ files using trim galore. Output in
trimmed_fastqfolder. - Align with STAR. Output in
STAR_alignment. - Filter SAM files, hardcoded to retain reads with mapping quality >=30 and non-multimappers (
-q 30 -F 4). Output infiltered_sam. - Remove duplicates—reads with the exact same UMI, barcode and genomic position. Output in
rmdup_sam. - Repeat duplicate removal–this time include with up to 1 mismatch in barcode+UMI sequence. Output in
rmdup_sam_2. - Splits the deduplicated SAM file into one file per valid barcode, at a given read cut-off. Output in
sam_splitted. - Generate gene/exon counts using HTSeq. Output in
report/human_mouse_gene_count. - [Outside of pipeline] Combine exonic and intronic reads and produce a count matrix across all cell barcodes and genes.
sci3_primer_sequences_plate.xls: primer sequences for reverse transcription, ligation, and PCR.
sci3_main.sh: main processing script for sci-RNA-seq3.
script_folder: folder for sub-scripts called by sci3_main.sh.
gene_count_processing_sciRNAseq.R: R script for processing the gene count data - the function of “sciRNAseq_gene_count_summary” accepts the gene count folder and then return a cell annotation data frame, a gene annotation data frame and a gene count sparse matrix. The output can be used as input to commonly used single cell RNA-seq analysis packages.