Enforce barcode match stringency and read quality in demultiplexed MiSeq reads.
Barcode mis-assignment represents a significant problem for ultra sensitive assays that strongly interpret the presence of very low prevalence reads within a specimen. In addition, mis-assignment of reads between specimens can create the appearance of template contamination in negative controls.
It turns out that barcode read quality is a predictor of barcode mis-assignment. Wright and Vetsigian (2016) (https://dx.doi.org/10.1186%2Fs12864-016-3217-x) showed that an average barcode quality score threshold of 26 prevented most read mis-assignment on the Illumina platform.
In addition, the onboard software used for demultiplexing on the Illumina MiSeq cannot be configured to enforce exact barcode matches. As a result, a minority of reads (up to about 5% in my tests) are assigned to a specimen on the basis of a partial barcode match. Somewhat anecdotally, these less-than-exact matches appear to have a higher likelihood of mis-assignment as well, but the effect is dependent on barcode sequence and which combinations of barcodes are used in the same library.
This package provides the barcodecop command that uses the index
reads to determine the most prevalent barcode sequence, and removes
reads without exact barcode matches from an accompanying fastq
file. It also filters reads based on average barcode quality score.
Command line arguments:
usage: barcodecop [-h] [-o OUTFILE] [--snifflimit N] [--head N] [--invert] [-q] [-V] (-c | -f file.fastq[.bz2|.gz]) [--match-filter]
[--min-pct-assignment PERCENT] [--strict] [-b FILE] [-C FILE] [--allow-empty] [--qual-filter] [-p MIN_QUAL] [--encoding {phred}]
file.fastq[.bz2|.gz] [file.fastq[.bz2|.gz] ...]
Filter fastq files by enforcing exact barcode match and quality score.
Input and output files may be compressed as indicated by a .bz2 or .gz
suffix.
positional arguments:
file.fastq[.bz2|.gz] one or two files containing index reads in fastq format
options:
-h, --help show this help message and exit
-o OUTFILE, --outfile OUTFILE
output fastq
--snifflimit N read no more than N records from the index file [10000]
--head N limit the output file to N records
--invert include only sequences failing filtering criteria
-q, --quiet minimize messages to stderr
-V, --version Print the version number and exit
-c, --show-counts tabulate barcode counts and exit
-f file.fastq[.bz2|.gz], --fastq file.fastq[.bz2|.gz]
reads to filter in fastq format
Barcode matching options:
--match-filter filter reads based on exact match to most common barcode [default: no match filter]
--min-pct-assignment PERCENT
warn (or fail with an error; see --strict) if the most common barcode represents less than PERCENT of the total [90.0]
--strict fail if conditions of --min-pct-assignment are not met
-b FILE, --barcode-counts FILE
tabulate barcode counts and store as a CSV
-C FILE, --read-counts FILE
tabulate read counts and store as a CSV
--allow-empty accept fastq files which contain zero length seqs/qual strings[default: do not allow empty seqs in fastq files]
Barcode quality filtering options:
--qual-filter filter reads based on minimum index quality [default: no quality filter]
-p MIN_QUAL, --min-qual MIN_QUAL
reject seqs with mean barcode quality score less than this value; for dual index, both barcodes must meet the threshold [26]
--encoding {phred} quality score encoding; see https://en.wikipedia.org/wiki/FASTQ_format [phred]
Install from PyPi using pip:
pip install barcodecop
Or install directly from the git repository:
pip install git+https://github.com/nhoffman/barcodecop.git
Run all tests like this:
python setup.py test