DNSS2

Deep learning architectures for protein secondary structure prediction (version 2). Both 3-class and 8-class prediction supported!

Test Environment

Red Hat Enterprise Linux Server release 6.4 (Santiago), perl 5, version 16, subversion 3 (v5.16.3)

Installation Steps

(A) Download and Unzip DNSS2 source package

Create a working directory called 'DNSS2' where all scripts, programs and databases will reside:

Download the DNSS2 code:

cd ~/
git clone https://github.com/multicom-toolbox/DNSS2.git
cd DNSS2

(B) Install tensorflow, Keras, and h5py and Update keras.json

(a) Create python2 virtual environment (if not installed)

virtualenv -p /usr/bin/python2  ./python_virtualenv_DNSS2
source ./python_virtualenv_DNSS2/bin/activate
pip install --upgrade pip

(b) Install python packages:

pip install -r requirements.txt

(f) Add the entry [“image_dim_ordering": "tf”,] to your keras..json file at ~/.keras/keras.json. After the update, your keras.json should look like the one below:

{
    "epsilon": 1e-07,
    "floatx": "float32",
    "image_dim_ordering":"tf",
    "image_data_format": "channels_last",
    "backend": "tensorflow"
}

(C) Download programs

mkdir programs
cd programs
wget https://ftp.ncbi.nih.gov/blast/executables/blast+/2.2.25/ncbi-blast-2.2.25+-x64-linux.tar.gz
tar -zxf ncbi-blast-2.2.25+-x64-linux.tar.gz


wget https://github.com/soedinglab/hh-suite/releases/download/v3.0-beta.3/hhsuite-3.0-beta.3-Linux.tar.gz
tar -zxf hhsuite-3.0-beta.3-Linux.tar.gz

(D) Download database

• DNSS2 requires non-redundent sequence database formated by blast(i.e., uniref90) and hhblits database(i.e., uniclust30_2017_10). If the two databases haven't been downloaded and formated, please try following steps:

mkdir database
cd database
mkdir uniref90
cd uniref90
wget ftp://ftp.uniprot.org/pub/databases/uniprot/uniref/uniref90/uniref90.fasta.gz
gzip -d uniref90.fasta.gz
../../programs/ncbi-blast-2.2.25+/bin/makeblastdb -in  uniref90.fasta

cd database
mkdir uniclust30_2017_10
cd uniclust30_2017_10
wget http://wwwuser.gwdg.de/~compbiol/uniclust/2017_10/uniclust30_2017_10_hhsuite.tar.gz

(E) Configuration

• Update the database paths for the following variables in the script 'configure.pl'

$uniref90db = 'your_db_path/uniref90/uniref90.fasta';
$uniclust30db = 'your_db_path/uniclust30_2017_10/uniclust30_2017_10';

(b) run 'configure.pl'

perl configure.pl

(F) Test

• There are two ways to indicate the protein to predict:

(1) Predict from protein file:

   Usage:
   $ perl run_DNSS2.pl -seq <file name>.fasta -file -out <output folder>

   Example:
   $ source ~/python_virtualenv_DNSS2/bin/activate
   $ mkdir -p ./output/2SN3-A
   $ perl run_DNSS2.pl -seq test/2SN3-A.fasta -file -out ./output/2SN3-A

[07/10/2020 23:19:49]  Generating Testing File...
[07/10/2020 23:19:49]  Accessing model deepss_1dconv
[07/10/2020 23:20:32]  Accessing model deepss_1dCRMN
[07/10/2020 23:21:05]  Accessing model deepss_1dFrac
[07/10/2020 23:21:35]  Accessing model deepss_1dInception
[07/10/2020 23:22:03]  Accessing model deepss_1dRCNN
[07/10/2020 23:22:29]  Accessing model deepss_1dResnet
[07/10/2020 23:22:54]  Ensemble predictions ...
[07/10/2020 23:25:54]  2SN3-A prediction (3-class) successful! Saved to ./output/2SN3-A/2SN3-A.ss8.dnss and ./output/2SN3-A/2SN3-A.ss8.vdnss
[07/10/2020 23:22:55]  Accessing model deepss_1dconv
[07/10/2020 23:23:19]  Accessing model deepss_1dCRMN
[07/10/2020 23:23:44]  Accessing model deepss_1dFrac
[07/10/2020 23:24:09]  Accessing model deepss_1dInception
[07/10/2020 23:24:35]  Accessing model deepss_1dRCNN
[07/10/2020 23:25:00]  Accessing model deepss_1dResnet
[07/10/2020 23:25:25]  Ensemble predictions ...
[07/10/2020 23:25:25]  2SN3-A prediction (8-class) successful! Saved to ./output/2SN3-A/2SN3-A.ss8.dnss and ./output/2SN3-A/2SN3-A.ss8.vdnss
[07/10/2020 23:25:25]  100% complete!

(2) Predicting multiple proteins:

   Usage:
   $ perl run_DNSS2.pl -indir <input directory> -out <output directory>

   Example:
   $ source ~/python_virtualenv_DNSS2/bin/activate
   $ mkdir ./output/
   $ perl run_DNSS2.pl -indir ./test/ -out ./output/