MGT-local

Install a local blank version of the MGT database

The MGT database and associated scripts are presented here in a format that should make them installable on a local machine. The MGT was written as a website and associated database so the setup of a local version is more complex than a normal command line program.

Installation:

1. input file generation

1. reference genome (fasta format)
2. lociLocationsInRef.txt - a tab delimited file with a row for each locus with columns:
   1. locus id
   2. reference genome start
   3. reference genome end
   4. + or - strand
   5. chromosome number (1 for all loci for most bacteria)
3. folder named "Schemes" - with one file for each MGT level named MGTN_gene_accessions.txt where N is the level number. Each file lists the loci to be used in each level.
4. setupPath.py - a file with variables for file paths to setup a new database.

examples are in the setup/example_inputs folder

2. install dependencies

1. install postgres
2. miniconda or mamba(recommended) for environment management
3. create an environment within conda/mamba using the included yaml file (/setup/mgt_conda_env.yaml). Default name is mgtenv.

   conda env create --file=/path/to/MGT-local/setup/mgt_conda_env.yaml
   or
   mamba env create --file=/path/to/MGT-local/setup/mgt_conda_env.yaml
   

3. modify settings

In the /Mgt/Mgt/Mgt folder find the settings_template.py file and rename any lines with #CHANGE comments as per comment instructions. You can also make a copy of this file and update the changes in the copy. Clawclip is an example of what the changes would look like.

4. run /setup/setup_new_database.ssh

use the command as follows from the setup directory: ./setup_new_database.ssh example_inputs.setupPath in command line use postgres password when prompted

5. access local mgt database site

run python manage.py runserver --settings Mgt.settings_template and access the website locally using host in settings file (http://localhost:8000/ by default)

6. typing isolates

Typing of isolates is in three stages:

1. Run reads_to_alleles.py script to generate an alleles file from reads or genomes following the readme in the /MGT-local/Mgt/Mgt/MGT_processing/Reads2MGTAlleles folder

2. Upload the alleles files along with associated metadata to the local site (via a project page)

3. run the /MGT-local/Mgt/Mgt/Scripts/cron_pipeline.py script to call final alleles and MGT types

   conda activate mgt_env
   
   cd /path/to/MGT-local/Mgt/Mgt/Scripts
   
   python cron_pipeline.py -s /path/to/settings_file.py -d Blankdb --allele_to_db --local