pydemux 1.0.1

Python package for demultiplexing raw sequencing reads in SAM/BAM format

pydemux

An easy to use python tool for fast demultiplexing of SAM/BAM formated raw sequence data

Installation

pip install pydemux

or clone the repository

git clone [e-mail]:dmalzl/pydemux.git

and run

cd pydemux
pip install .

verify the installation by typing

pydemux -h

Basic usage

pydemux can be used to demultiplex sequence data in SAM/BAM format for either single or paired-end reads (in case of paired-end the reads have to be interleaved). Basic command for this are listed below (you can test these using the files in the data directory)

pydemux single -b single_barcodes.tsv -o demux/ -s demux_stats.tsv single.bam

pydemux paired -b paired_barcodes.tsv -o demux/ -s demux_stats.tsv paired.bam

Note that for writing the files to a specific output directory the / needs to be included for the -o/--output_prefix commandline argument or they will be written to the current work directory. In order to compress the output one can also add the -gz/--gzip commandline argument to the command.

Changing looked up SAMfile tags from which the barcodes are read

By default the algorithm looks for the barcode in the BC tag of each read in case of pydemux single and in the BC and B2 tag of the paired-end reads in case of pydemux paired. This can be changed using either the -t/--bctag or -t1/--bc1tag and/or -t2/--bc2tag for pydemux single or pydemux paired respectively.

Speeding up processing

In order to speed up demultiplexing the algorithm can be run concurrently using the -p/--processes argument.

Optimizing yield

As with every sequencing-based data type, the barcodes are also prone to include sequencing errors. In order to optimize the read yield for each sample the algorithm allows for a given number of mismatches between the true and the sequenced barcodes which can be set with -m/--mismatches. By default only exact matches will be assigned. If you want to allow for one or more mismatches please make sure that the number of allowed mismatches does not exceed half of the minimum pairwise Hamming distance of all true barcodes minus 1 (i.e. min(pairwise_hamming_distance(true_barcodes)) // 2 - 1) since otherwise reads might be wrongly assigned

Barcode file format

The barcode file is a simple tab-separated file containing one sample and its associated barcodes per line. In case of pydemux single each line consists of one barcode and one sample name. e.g.

barcode	sample_name
ACGTCGTA	sample_1

In case of pydemux paired the file contains two barcodes and one sample name per line. e.g.

barcode1	barcode2	sample_name
ACGTCGTA	CGTAGGAT	sample_1