NG-STAR (Neisseria gonorrhoeae Sequence Typing for Antimicrobial Resistance) is a typing scheme published by Walter Demczuk and colleagues (J Clin Microbiol 2017 55:1454–1468. https://doi.org/10.1128/JCM.00100-17) that targets genes associated to cephalosporin, macrolides and fluoroquinolones resistance:
penA, mtrR, porB, ponA, gyrA, parC and 23S rRNA
The NG-STAR database and web application is hosted by the Public Health Agency of Canada, National Microbiology Laboratory (walter.demczuk@canada.ca) and can be accessed from: https://ngstar.canada.ca. More information on the scheme can be found in that same webpage and their paper cited above.
pyngSTar is a python3 script to do NG-STAR typing in Neisseria gonorrhoeae genome assemblies. To look for exact matches, it can be run on a slow(er) or a fast mode (-f) if you have a large collection.
- Fast mode (-f, recommended): uses a python module implementing the Aho-Corasick algorithm for multi-pattern string search (https://github.com/WojciechMula/pyahocorasick/).
- Slow(er) mode: loops through the keys of a dictionary containing the forward and reverse complementary of the alleles of the seven genes.
The main script is called pyngSTar.py and is accompanied by two modules:
- pyngSTar_functions.py: contains basic functions for reading input files, assigning allele numbers (slow mode), reporting profiles, blasting and printing new allele sequences.
- pyngSTar_AhoCorasick.py: imports the pyahocorasick module for fast multi-pattern string search. This is provided separately in case anyone has problems with installing the library (unlikely) and decide to use the slow mode.
If an exact match is not found, blastn is called to report the closest matches and the new allele is printed to an output file if requested with -a. Closest matches are marked in the profiles table with an asterisk.
NOTE: If more than 3 alleles are equally close, only 3 are shown in the table. In this case, something may be wrong, i.e. the whole sequence of the allele is not found and only part of it is being considered, so it is good to take a look at the length of the sequences of the potential new alleles requested with -a.
Profiles not found in the local database are marked in the table as 'NEW'.
From the seven genes in the scheme, we only expect to have more than one copy in the case of 23S (4 copies).
pyngSTar will detect if there is more than one allele and output them separated by an underscore '_'. In this case, the final ST will appear as 'NEW' and the user will have to work out if any of those alleles are part of an existing profile.
- os, subprocess, argparse
- pickle
- pandas
- SeqIO (biopython)
- pyfaidx
- pyahocorasick (optional)
Use pip3 install and a list of the missing modules:
pip3 install pandas pyfaidx biopython pyahocorasick
If you need some of the others just use pip3 install also on those.
Then I recommend to create a python3 virtual environment following these steps:
Create and activate a virtual environment (venv or any other name):
virtualenv venv
source venv/bin/activate
Install dependencies inside the environment:
pip3 install pandas pyfaidx biopython pyahocorasick
Run pyngSTar:
python3 pyngSTar.py -f -a -i /path/to/assemblies/*.fa -p pyngSTarDB/
Exit the virtual environment:
deactivate
IMPORTANT: You only need to create the virtual environment and install dependencies once, afterwards just activate it, run pyngSTar and then deactivate when it’s finished:
# activate virtualenv
source venv/bin/activate
# run pyngSTar
# exit virtualenv
deactivate
usage: pyngSTar.py [options]
Get NG-STAR types from N. gonorrhoeae assemblies
optional arguments:
-h, --help show this help message and exit
-i INPUT [INPUT ...], --input INPUT [INPUT ...]
Assembly files
-p PATH, --path PATH Path to database containing alleles and profiles
-f, --fast Uses the Aho-Corasick algorithm for fast searching
(optional)
-a, --alleles_out Print fasta files with new alleles (optional)
-o OUT_FILENAME, --out_filename OUT_FILENAME
Name of file to print output to (optional, default:
screen output)
| Option | Description |
|---|---|
| -i | list of assembly files (* can be used to input multiple files in a directory). |
| -p | path to a folder containing the database files. This folder must contain 9 files. |
| -f | fast allele searching option using the Aho-Corasick algorithm (optional but recommended). |
| -a | print new allele sequences if found (optional). |
| -o | output filename (default is output to screen). |
A copy of the database (downloaded on July 2019 from https://ngstar.canada.ca) folder is available in pyngSTarDB.tar.gz and contains:
- <gene>_alleles.fasta: 7 allele sequence files with this structure.
- ngstar_profiles.tab: 1 file containing the profiles.
- pyngSTar_alleles_AC.pkl: 1 pickle file containing the preloaded dictionary of allele sequences and numbers as well as the automaton object required by the fast searching algorithm.
To run pyngSTar in a fast mode (-f) and requesting new alleles to be printed as fasta files (-a):
python3 pyngSTar.py -f -a -i /path/to/assemblies/*.fa -p pyngSTarDB/
Add -o <filename> if you want the output to go to a file instead of the screen.
strain ngSTar penA mtrR porB ponA gyrA parC 23S
10356_1#33.fa NEW 19.001 25|86* 100 1 100 7 100
10625_6#40.fa NEW 14.001* 54 100 1 1 18 100
10625_6#8.fa 127 5.002 1 8 1 1 3 100
11792_4#61.fa 356 2.001 10 100 100 1 18 100
This will produce two fasta files containing the new alleles found for penA and mtrR. The header of the sequences indicate have the gene name, the contig name where the gene was found and the start and end coordinates in that contig where the sequence was pulled out from:
>mtrR_.10356_1_33.8_29704:30403
TTGCATGGTTACAAAGTCTTTTTTATAATCCGCCCTCATCAAACCGACCCGAAACGAAACCGCCA...
For the 419 N. gonorrhoeae strains from Sánchez-Busó et al. 2019 (https://www.nature.com/articles/s41564-019-0501-y), pyngSTar on the fast mode took just 1min 31s!!
Go to https://ngstar.canada.ca and manually download the profiles table and the 7 fasta files of alleles. Copy/paste the profiles table from the Excel file to a plain text file and call it 'ngstar_profiles.tab' (replace "Sequence Type" by "ST" in the header). Move the 8 files (table in text format and fasta files) to the database folder. Recreate the pickle file (pyngSTar_alleles_AC.pkl), which will boost the speed of subsequent runs on the --fast mode, by running:
python3 pyngSTar.py -p pyngSTarDB -u
Same as with NG-MASTER (https://doi.org/10.1099/mgen.0.000076), each typing scheme has its own particularities!!
If you use it and it crashes, or you have an idea to improve it, i'd love to hear from you (leonor.sanchezbuso@bdi.ox.ac.uk).
Please, if you use this script for a publication, cite the original work by Walter Demczuk and colleagues https://doi.org/10.1128/JCM.00100-17 and a link to this github page :)