synthbar

tl;dr

synthbar adds a synthetic cell barcode to each read in a provided FASTQ.

Install and Build

git clone git@github.com:jamorrison/synthbar.git
cd synthbar
make

Overview

Most single cell RNA-seq protocols include a cell barcode at the beginning of each read to distinguish which cell the read came from. This barcode is followed by a unique molecular index (UMI), which allows for tracking individual RNA fragments and correct deduplication of RNA sequencing data. Finally, a short linker sequence is required to connect the cell barcode and UMI to the cDNA insert that will be sequenced.

While the majority of single cell protocols include both the cell barcode and the UMI, some protocols (particularly plate based single cell protocols and even some bulk protocols) utilize a UMI without a cell barcode (see Note 1 below). For historical reasons, most tools for analyzing RNA-seq data with UMIs require a cell barcode before the UMI and can't handle reads without barcodes.

To overcome this hurdle, synthbar prepends a 7-base cell barcode (CATATAC) to the sequence string of each read in the FASTQ, as well as a matching 7-base quality score (IIIIIII) in the quality string. If the user would rather use a different barcode, the --barcode option allows for a user-defined barcode. In this case, a string of I's of length equal to the requested barcode will be prepended to the quality string. If desired, the linker sequence can be removed as well, leaving just the barcode, UMI, and cDNA sequence.

Note 1: With respect to the plate based single cell protocols, individual cells are sorted into each well in a plate and then an index is added to distinguish cells for demultiplexing after sequencing (similar to distinguishing samples on your standard Illumina sequencer). In the case of bulk protocols, a "cell barcode" by itself doesn't make sense as there are many cells within the sample. However, even in bulk protocols, UMIs can serve a scientific purpose. In order to properly account for the UMIs in these protocols, they should be processed in a similar manner to the various single cell protocols with UMIs.

Usage

Usage: synthbar [options] <FASTQ with UMIs>

Output options:
    -o, --output STR           name of output file [stdout]
Processing Options:
    -b, --barcode STR          barcode to prepend to each read [CATATAC]
    -U, --umi-first            add barcode to read after the UMI [off]
    -r, --remove-linker        remove linker from read [not removed]
    -l, --linker-length INT    length of linker to remove [6]
    -u, --umi-length INT       length of UMI before linker [8]
    -h, --help                 print usage and exit
        --version              print version and exit

Note 1: Input FASTQ can be gzip compressed or uncompressed

Option	Input	Description
-o, --output	string	name of output file (defaults to stdout), does not write gzip'd files
-b, --barcode	string	barcode to add instead of CATATAC (does not check if composed of ATCG's)
-U, --umi-first	-	place the barcode after the UMI in the new read
-r, --remove-linker	-	remove linker sequence from read (not removed by default)
-l, --linker-length	integer (>= 0)	length of linker to remove (default is 6), not used if -r not provided
-u, --umi-length	integer (>= 0)	length of UMI before linker (default is 8), not used if -r not provided
-h, --help	-	print usage and exit
--version	-	print version and exit

Note, for protocols with no linking sequence, it is suggested to ignore the linker-related options, as this will ensure everything is written after the UMI and eliminate the potential for inadvertently removing cDNA sequence.

Read Structure

In / Out	Linker?	UMI First?	Remove Linker?	Structure
Input	Yes	NA	NA	( UMI ) + ( LINKER ) + ( cDNA )
Input	No	NA	NA	( UMI ) + ( cDNA )
Output	Yes	No	No	( BARCODE ) + ( UMI ) + ( LINKER ) + ( cDNA )
Output	Yes	No	Yes	( BARCODE ) + ( UMI ) + ( cDNA )
Output	Yes	Yes	No	( UMI ) + ( BARCODE ) + ( LINKER ) + ( cDNA )
Output	Yes	Yes	Yes	( UMI ) + ( BARCODE ) + ( cDNA )
Output	No	No	No	( BARCODE ) + ( UMI ) + ( cDNA )
Output	No	Yes	No	( UMI ) + ( BARCODE ) + ( cDNA )

Acknowledgments

• synthbar uses two utilities from klib, kseq (for handling FASTQs) and kstring (for writing updated reads).

Citation

At this time, there is no plan to publish synthbar. However, if you would kindly cite the GitHub page whenever you use synthbar in your analysis, the reference would be greatly appreciated.