Skip to main content

ContigNet, a deep learning based phage-host interaction prediction tool

Project description

ContigNet: Phage-bacteria contig interaction prediction with convolutional neural network

Tests

The paper has been published at https://doi.org/10.1093/bioinformatics/btac239

Version: 1.0.1

Authors: Tianqi Tang, Shengwei Hou, Jed Fuhrman, Fengzhu Sun

Maintainer: Tianqi Tang tianqit@usc.edu

Description

This is the repository containing the software ContigNet and related scripts for the paper "Phage-bacteria contig interaction prediction with convolutional neural network".

ContigNet is a deep learning based software for phage-host contig interaction prediction. Traditional methods can work on contigs however the performance is poor. Existing Deep learning based methods are not able to solve the particular question regarding interaction prediction between two contigs.

Installation

To use the software, download and enter the repository by

git clone https://github.com/tianqitang1/ContigNet
cd ContigNet

To install required dependencies a Anaconda or Miniconda installation is recommended for managing virtual environments. After a conda distribution is installed, create and activate a conda virtual environment with the following commands

conda create --name ContigNet
conda activate ContigNet
pip install .

Usage

usage: ContigNet [-h] [--host_dir HOST_DIR] [--virus_dir VIRUS_DIR]
                    [--output, -o OUTPUT] [--cpu]

ContigNet, a deep learning based phage-host interaction prediction tool

optional arguments:
  -h, --help            show this help message and exit
  --host_dir HOST_DIR   Directory containing host contig sequences in fasta
                        format (default: demo/host_fasta)
  --virus_dir VIRUS_DIR
                        Directory containing virus contig sequences in fasta
                        format (default: demo/virus_fasta)
  --output, -o OUTPUT   Path to output file (default: result.csv)
  --cpu                 Force using CPU if specified (default: False)

Examples

Test new contigs

Suppose the phage and host sequences are stored in phage and host directories respectively, running

ContigNet --host_dir host --virus_dir phage

and the likelihood of each phage interacting with each host will be output to result.csv.

For Windows machine, run

python -m ContigNet --host_dir host --virus_dir phage

Paper related

Browse training directory for the instructions of running the training and testing process for the paper.

Copyright and License Information

Copyright (C) 2021 University of Southern California

Authors: Tianqi Tang, Shengwei Hou, Jed Fuhrman, Fengzhu Sun

This program is available under the terms of USC-RL v1.0.

Commercial users should contact Dr. Sun at fsun@usc.edu, copyright at the University of Southern California.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

ContigNet-1.0.1.post3.tar.gz (5.1 MB view hashes)

Uploaded Source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page