dreyceyalbin/phagescanner repository overview

⁠PhageScanner

PhageScanner is a command line tool for identifying phage virion proteins (PVPs) using metagenomic sequencing data as input. For comprehensive information about installation and usage, please visit the PhageScanner Wiki⁠.

⁠Installing using Docker (Windows, Mac and Linux)

The easiest approach to using PhageScanner is to use Docker. Docker allows for PhageScanner to be usable on Windows and removes the need to install the command line tool dependencies. Follow the directions to install docker⁠. For Windows, we used WSL2 to install docker (instead of Hyper-V), but both should work as intended.

⁠Using the Docker image host on DockerHub

• Pull down the docker image from DockerHub

docker pull dreyceyalbin/phagescanner

• Test that the help message prints

docker run --rm dreyceyalbin/phagescanner --help

⁠Building Docker image locally

The docker image can be built locally to allow for more flexiblity. There are two steps involved in this process:

• Navigate to the Docker/ directory and run:

docker build -t dreyceyalbin/phagescanner .

• Test that the help message prints

docker run --rm dreyceyalbin/phagescanner --help

⁠Pipeline Usage

There are three fundamental pipelines in the PhageScanner tool. Each of these pipelines feeds into the next: (1) Download the training dataset, (2) Training the machine learning models, (3) Using the models to annotate genomes and metagenomics datasets. Each pipelines is configurable to allow end-users extreme flexibility in creating new models to predict new variations of protein classes (ex. "Toxic Protein", "Phage Virion Protein", "Lysogenic"). Each example list below should be ran from the root directory if running the commands "as-is".

1. Build the database
   • Basic usage

   python phagescanner.py database [-h] -c CONFIG -o OUT [--cdhit_path CDHIT_PATH (Default: 'cdihit')] [-v VERBOSITY]
   

   Copy
   • Example (multiclass pvps)

   python phagescanner.py database -c configs/multiclass_config.yaml -o ./multiclass_database/ -v info
   

   Copy
   • Example using Docker (multiclass pvps)

   docker run --rm \
       -v "$(pwd)/configs:/app/configs" \
       -v "$(pwd)/multiclass_database:/app/multiclass_database" \
       dreyceyalbin/phagescanner database -c /app/configs/multiclass_config.yaml -o /app/multiclass_database/ -v info
   

   Copy
2. Training and Test ML models
   • Basic usage

   python phagescanner.py train [-h] -c CONFIG -o OUT -db DATABASE_CSV_PATH [-v VERBOSITY]
   

   Copy
   • Example (multiclass pvps)

   python phagescanner.py train -c configs/multiclass_config.yaml -o training_output --database_csv_path ./multiclass_database/ -v debug
   

   Copy
   • Example using Docker (multiclass pvps)

   docker run --rm \
       -v "$(pwd)/configs:/app/configs" \
       -v "$(pwd)/multiclass_database:/app/multiclass_database" \
       -v "$(pwd)/training_output:/app/training_output" \
       dreyceyalbin/phagescanner train -c /app/configs/multiclass_config.yaml -o /app/training_output --database_csv_path /app/multiclass_database/ -v debug
   

   Copy
3. Run on metagenomic data, genomes or proteins
   • Basic usage

   python phagescanner.py predict [-h] -i INPUT -t TYPE ("reads", "genome", or "protein") -c CONFIG -o training_output -n NAME -tdir TRAINING_OUTPUT
                               [--megahit_path MEGAHIT_PATH (Default: 'megahit')] [--phanotate_path PHANOTATE_PATH (Default: 'phanotate.py')]
                               [--probability_threshold PROBABILITY_THRESHOLD] [-v VERBOSITY]
   

   Copy
   • Example (genomes; though sequencing reads and proteins can be used as input)

   python phagescanner.py predict -c configs/multiclass_config.yaml  -t "genome" -o prediction_output -n "genomes" -i examples/GCF_000912975.1_ViralProj227117_genomic.fna -v debug
   

   Copy
   • Example using Docker (genomes)

   docker run --rm \
       -v "$(pwd)/configs:/app/configs" \
       -v "$(pwd)/examples:/app/examples" \
       -v "$(pwd)/prediction_output:/app/prediction_output" \
       dreyceyalbin/phagescanner predict -c /app/configs/multiclass_config.yaml -t "genome" -o /app/prediction_output -n "genomes" -i /app/examples/GCF_000912975.1_ViralProj227117_genomic.fna -v debug
   

   Copy