################################################################################
Name:
    PPVED - Plant Protein Variation Effect Detector

Version:
    1.0

Author:
    Xiangjian Gou (Sichuan Agricultural University)

Date:
    2021/01/20

Function:
    PPVED predicts the functional effect of single amino acid substitution in
    plants using the ensemble of XGBoost models.

Contact:
    Xiangjian Gou
    Sichuan Agricultural University
    xjgou@stu.sicau.edu.cn

    Yanli Lu
    Sichuan Agricultural University
    luyanli@sicau.edu.cn
################################################################################


Next, this document will teach you how to download, install and use PPVED.


1. Download

You can download PPVED from http://www.ppved.org.cn/packages/PPVED-1.0.tar.gz

2. Install

PPVED was installed and tested on a 64 bit CentOS Linux server. You should need at
least 60Gb of disk space to complete the installation of PPVED. The installation of
PPVED is divided into the following five steps:
    (1) Decompress PPVED.
            tar zxvf PPVED-1.0.tar.gz
            cd PPVED-1.0

    (2) Check whether Perl and R languages exist.
            perl -v
            Rscript --version

    (3) Install dependent seven prediction software.
            Read the file 'install/install_dependent_software.txt' to learn how to
            install the seven prediction software.

            When you finish the installation, please modify the content of the file
            'required_software.txt' to suit your system.

    (4) Install dependent R package 'xgboost'.
            First, we assume that you intend to install the R package in the path
            '/datadisk/PPVED-1.0/rpackage':
                cd rpackage
                echo "export R_LIBS=/datadisk/PPVED-1.0/rpackage:\$R_LIBS" >> ~/.bashrc
                source ~/.bashrc

            Before installing the package 'xgboost', you may need to install several
            dependent packages (e.g. data.table, magrittr and stringi), of course,
            these packages may already exist in your system:
                R CMD INSTALL -l /datadisk/PPVED-1.0/rpackage data.table_1.13.6.tar.gz
                R CMD INSTALL -l /datadisk/PPVED-1.0/rpackage magrittr_2.0.1.tar.gz
                tar zxvf stringi.tar.gz
                #note: here, package 'stringi' is not installed through source code.

            Finally, install the package 'xgboost':
                R CMD INSTALL -l /datadisk/PPVED-1.0/rpackage xgboost_0.90.0.2.tar.gz

    (5) Set the search path of shared library files.
            echo "export LD_LIBRARY_PATH=.:\$LD_LIBRARY_PATH" >> ~/.bashrc
            source ~/.bashrc

    Congratulations, the installation is complete. Please do not move any file of PPVED !

3. Test

    (1) Change the current working directory to the subdirectory 'example', and then:
            perl ../PPVED.pl -i example.input -o example.test
            #The output file 'example.test' should be identical to the provided file 'example.output'.

    (2) You can see the format of input file by using the following command:
            perl PPVED.pl -e

    (3) You can also see the help document of PPVED by using the following command:
            perl PPVED.pl -h

4. Format of output

The local version of PPVED supports batch calculation, and the output file contains
a total of five columns:
    col1: the single amino acid substitution number
    col2: the name of protein
    col3: the single amino acid substitution
    col4: the predicted probability score, the value is between 0-1
    col5: the predicted binary classification (functional: score >= 0.5; neutral: score < 0.5)