DeePhage: a tool for identifying temperate phage-derived and virulent phage-derived sequence in metavirome data using deep learning

• Introduction
• Version
• Requirements
• Installation
• Usage
• Output
• Citation
• Contact

Introduction

DeePhage is designed to identify metavirome sequences as temperate phage-derived and virulent phage-derived sequences. The program calculate a score reflecting the likelihood of each input fragment as temperate phage-derived and virulent phage-derived sequences. DeePhage can run either on the virtual machine or physical host. For non-computer professionals, we recommend running the virtual machine version of DeePhage on local PC. In this way, users do not need to install any dependency package. If GPU is available, you can also choose to run the physical host version. This version can automatically speed up with GPU and is more suitable to handle large scale data. The program with original data could be downloaded from data.zip.

Version

• DeePhage 1.0 (Tested on Ubuntu 16.04)

Requirements

---

1. To run the physical host version of DeePhage, you need to install:

• Python 3.6.7

• numpy 1.16.4

• h5py 2.9.0

• TensorFlow 1.4.0

• Keras 2.1.3

• MATLAB Component Runtime (MCR) R2018a or MATLAB R2018a

  Note: (1) DeePhage should be run under Linux operating system. (2) For compatibility, we recommend installing the tools with the similar version as described above. (3) If GPU is available in your machine, we recommend installing a GPU version of the TensorFlow to speed up the program. (4) DeePhage can be run with either an executable file or a MATLAB script. If you run DeePhage through the executable file, you need to install the MCR (for free) while MATLAB is not necessary. If you run DeePhage through the MATLAB script, MATLAB is required.

2. If you are non-computer professionals who unfamiliar with the Linux operating system, we recommend using the virtual machine of DeePhage. In this way, you do not need to install any dependant packages as mentioned above.

Installation

1. Physical host version

1.1 Prerequisites

First, please install numpy, h5py, TensorFlow and Keras according to their manuals. All of these are python packages, which can be installed with pip. If “pip” is not already installed in your machine, use the command “sudo apt-get install python-pip python-dev” to install “pip”. Here are example commands of installing the above python packages using “pip”.

pip install numpy
pip install h5py
pip install tensorflow==1.4.0  #CPU version
pip install tensorflow-gpu==1.4.0  #GPU version
pip install keras==2.1.3

If you are going to install a GPU version of the TensorFlow, specified NVIDIA software should be installed. See https://www.tensorflow.org/install/install_linux to know whether your machine can install TensorFlow with GPU support.

When running DeePhage through the executable file, MCR should be installed. See https://www.mathworks.com/help/compiler/install-the-matlab-runtime.html to install MCR. On the target computer, please append the following to your LD_LIBRARY_PATH environment variable according to the tips of MCR:

<MCR_installation_folder>/v94/runtime/glnxa64
<MCR_installation_folder>/v94/bin/glnxa64
<MCR_installation_folder>/v94/sys/os/glnxa64
<MCR_installation_folder>/v94/extern/bin/glnxa64

When running DeePhage through the MATLAB script, please see https://www.mathworks.com/support/ to install the MATLAB.

1.2 Install DeePhage using git

Clone DeePhage package

git clone https://github.com/shufangwu/DeePhage.git

Change directory to DeePhage:

cd DeePhage

The executable file and all scripts are under the folder

1.3 Install DeePhage from zipped file

DeePhage can also download as a zipped file:

wget http://cqb.pku.edu.cn/ZhuLab/DeePhage/DeePhage_v_1_0.zip

Unpack the zipped file:

unzip DeePhage_v_1_0.zip

Change directory to DeePhage

cd DeePhage_v_1_0

The executable file and all scripts are under the folder

2. Virtual machine version

The installation of the virtual machine is much easier. Please refer to Manual.pdf for a step by step guide with screenshot to see how to install the vertual machine.

Usage

1. Run DeePhage using executable file (in command line)

Please simply execute the command:

./DeePhage <input_file_folder>/input_file.fna <output_file_folder>/output_file.csv

The input file must be in fasta format containing the sequences to be identified. For example, users can use the file "example.fna" in the folder to test DeePhage by simply executing the command:

./DeePhage example.fna result.csv

2. Run DeePhage using MATLAB script (in MATLAB GUI)

Please execute the following command directly in MATLAB command window:

DeePhage('<input_file_folder>/input_file.fna','<output_file_folder>/output_file.csv')

For example, if you want to identify the sequences in "example.fna", please execute:

DeePhage('example.fna','result.csv')

Please remember to set the working path of MATLAB to DeePhage folder before running the programe.

3. Run DeePhage with specified cutoff

For each input sequence, DeePhage will output one scores (between 0 to 1), representing the lifestyle prediction score. The sequence with a score higher than 0.5 would be regarded as a virulent phage-derived fragment and the sequence with a score lower than 0.5 would be regarded as a temperate phage-derived fragment. Users can also specify a cutoff. In this way, a sequence with a score between (0.5-cutoff, 0.5+cutoff) will be labelled as "uncertain". In general, with a higher cutoff, the percentage of uncertain predictions will be higher while the remaining predictions will be more reliable. For example, if you want to get relatively reliable predictions in the file “example.fna”, you can take 0.3 as the cutoff. Please execute:

./DeePhage example.fna result.csv 0.7 (by executable file)
or
DeePhage('example.fna','result.csv','0.7') (by MATLAB script)

4. Run DeePhage in virtual machine

We recommend that non-computer professionals run DeePhage in this way. The virtual machine version of PPR-DeePhage run through the executable file (see item 1 above). Please refer to Manual.pdf.

Output

The output of DeePhage consists of four columns:

Header	Length	lifestyle_score	possible_lifestyle

The content in Header column is the same with the header of corresponding sequence in the input file. The possible lifestyle is the the possible lifestyle of the sequence. If DeePhage is executed under specified threshold, the sequence with the highest score lower than the threshold will be label as uncertain in possible_lifestyle columns.

Note: (1) The current version of DeePhage uses “comma-separated values (CSV)” as the format of the output file. Please use “.csv” as the extension of the output file. DeePhage will automatically add the “.csv” extension to the file name if the output file does not take “.csv” as its extension”. (2) If you want to run multiple tasks at the same time (either on physical host or virtual machine), please copy DeePhage package into different folders and run different tasks under different folders. Do not run different tasks under the same folder.

Citation

Wu, S., Fang, Z., Tan, J., Li, M., Wang, C., Guo, Q., ... & Zhu, H. (2021). DeePhage: distinguishing virulent and temperate phage-derived sequences in metavirome data with a deep learning approach. GigaScience, 10(9), giab056.

Contact

Any question, please do not hesitate to contact me: wu-shufang@pku.edu.cn