PRGminer: Deep neural network based Plant Resistance Gene Prediction

PRGminer classify Plant resistance genes in 8 classes: CNL, KIN, LYK, LECRK, RLP, RLK, TIR, TNL.

Input

The PRGminer server requires amino acid sequence(s) in fasta format, and does not support nucleic acid sequences.

Development Environment and Prerequisite

This source code was developed in Linux, and has been tested on Linux and OS X. The only prerequisite is to have Python 3.7 or above installed.

Installation

The installation of PRGminer can be done in two ways:
        
        1. Create a dedicated miniconda3 environment
        
            Download PRGminer from:
                
                https://bioinfo.usu.edu/PRGminer/download/PRGminer.tar.gz
        
            Download the Miniconda installer: 
                
                (https://docs.conda.io/en/latest/miniconda.html#linux-installers)
        
            Extract the downloaded file:
        
                tar -xvzf PRGminer.tar.gz
        
            cd PRGminer
            
            Create and activate a conda environment
        
                conda env create -f environment.yml
        
                conda activate PRGminer
        
                pip3 install .
        
        2. Intall using system Python3
        
            Download PRGminer from: 
        
            https://bioinfo.usu.edu/PRGminer/download/PRGminer.tar.gz
        
            Extract the downloaded file:
        
                tar -xvzf PRGminer.tar.gz
        
            cd PRGminer
        
                pip3 install .

        3. Intall using system Python3

            Clone PRGminer from: 
        
            git clone https://github.com/usubioinfo/PRGminer.git
        
            cd PRGminer
        
                pip3 install .
        

Input and Execution

PRGminer will be installed under the name 'PRGminer' with three possible arguments:
        
        -i, --fasta_file. Protein sequence input in FASTA format.
        -od, --output_dir. Output directory name.
        -l, --level. 
        

Output Files

The output directory will have three files in tab-delimted format:

1. Phase1_dnn_log.txt

   • 1st column: Protein ID.
   • 2nd column: Prediction.
   • 3rd-4th column: Probability for Rgenes and Non-Rgenes.

2. Phase2_dnn.log.txt

   • 1st column: Protein ID.
   • 2nd column: Prediction.
   • 3rd-14th column: Probability for each class.

Queries and Contact

Written by Naveen Duhan (naveen.duhan@usu.edu),

Kaundal Bioinformatics Lab, Utah State University,

Released under the terms of GNU General Public Licence v3

In case of technical problems (bugs etc.) please contact Naveen Duhan naveen.duhan@usu.edu.

For any Questions on the scientific aspects of the PRGminer method please contact:

Rakesh Kaundal, rkaundal@usu.edu

Naveen Duhan, naveen.duhan@usu.edu