pysarg

Python implementation of ARGs_OAP

Installation

• build from source (cmake, zlib and libpthread are required for building diamond and minimap2):

git clone https://github.com/xinehc/pysarg
cd pysarg
## use python3 if needed
python setup.py install

• the source code is also uploaded to pypi, try:

## make take a while for building the wheel, as 2 different versions of diamond need to be compiled
## use pip3 if needed
pip install pysarg

• pre-compiled conda packages (osx-64 or linux-64, python 3.6)

conda install -c xinehc pysarg

• only python=3.6 package has been uploaded, if python!=3.6, create a new conda environment

conda create -n pysarg -c xinehc pysarg python=3.6
source activate pysarg

Example

Two toy examples (100k paired-end reads, 100bp each) are provided in example/inputdir:

# git clone https://github.com/xinehc/pysarg
# mkdir -p pysarg/example/outputdir
pysarg stage_one -i pysarg/example/inputdir -o pysarg/example/outputdir
pysarg stage_two -i pysarg/example/outputdir/extracted.fasta -m pysarg/example/outputdir/metadata.txt -o pysarg/example/outputdir 

## LINUX only: add flag --original to get Ublastx_stageone's results (required to use the pre-compiled binaries in Ublastx_stageone)
pysarg stage_one -i pysarg/example/inputdir -o pysarg/example/outputdir --original
pysarg stage_two -i pysarg/example/outputdir/extracted.fasta -m pysarg/example/outputdir/metadata.txt -o pysarg/example/outputdir --original

If everything is ok, there should be four output files in example/outputdir

• metadata.txt

sample	read_length	read_number	16s_number	cell_number
STAS	100	200000	9.776536312849162	3.05292019025543
SWHAS104	100	200000	9.35754189944134	3.3635174193105737

• output.txt

sample	sequence	gene	gene_length	gene_type	gene_subtype	covered_length
STAS	STAS_30	gi|671541568|ref|WP_031525212.1|	648	macrolide-lincosamide-streptogramin	macrolide-lincosamide-streptogramin__macB	31
STAS	STAS_61	NP_840140	273	bacitracin	bacitracin__bacA	32
STAS	STAS_70	gi|764440891|ref|WP_044366757.1|	439	multidrug
...	...	...	...	...	...	...

• the above two tables can be merged on column sample and then used for normalizing the ARG counts and drawing PCA plots. One example is provided in notebook/normalize_sarg.ipynb
• extracted.fasta is the pre-filtered ARG-like sequences in stage-one, extracted.blast is the blastx result of extracted.fasta in stage-two.