Installation and requirements

Note: This tutorial describes how the eThread webserver was set up and how it works. The installation of eThread and its many dependencies requires lots of patience. If you have only a few targets, simply use the webserver. If you still want to use it locally, follow the steps below. Here, we will install everything in /usr/local/, if you want to use a different location, make the appropriate changes.

Perl modules

First, install the following Perl modules, which are available from CPAN:

AI::NaiveBayes1 http://search.cpan.org/~vlado/AI-NaiveBayes1-2.006/
Algorithm::NeedlemanWunsch http://search.cpan.org/~vbar/Algorithm-NeedlemanWunsch-0.03/
Bit::Vector http://search.cpan.org/~stbey/Bit-Vector-7.1/
Compress::Zlib http://search.cpan.org/~pmqs/IO-Compress-2.043/
File::Slurp http://search.cpan.org/~uri/File-Slurp-9999.19/
Math::MatrixReal http://search.cpan.org/~leto/Math-MatrixReal-2.09/
Math::Trig http://search.cpan.org/~zefram/Math-Complex-1.58/
Statistics::Descriptive http://search.cpan.org/~shlomif/Statistics-Descriptive-3.0203/
Uniq http://search.cpan.org/~syamal/Uniq-0.01/
YAML http://search.cpan.org/~ingy/YAML-0.77/

For instructions on how to install Per modules, see http://www.cpan.org/modules/INSTALL.html.

Installation

Download the latest version of eThread from here, unpack the tarball, go to ethread-1.0/src/ and run make:

[local]$ tar xzf ethread-1.0.tar.gz

[local]$ cd ethread-1.0/src/

[src]$ make

If there are no problems, the following files will show up in ethread-1.0/bin/:

[src]$ ls ../bin/

econtact
econtact_model
econv-compass
econv-csiblast
econv-fix1
econv-hhpred
econv-hmmer
econv-pftools
econv-pgenthreader
econv-samt2k
econv-sp3
econv-sparks
econv-threader
eextract
emodel
erank-modeller
erank-tasser
ethread
ethread_model

Next, build machine learning models:

[src]$ make models

Building these models may take a very long time, even 3-4 hours depending on how fast is your CPU. The reason for not distributing pre-compiled models is that we found some incompatibilities caused by different Perl module versions and across different architectures and operating systems. You can start making models in the background and continue with the installation. Also, note that this step requires around 3.3GB of memory, so make sure there is enough RAM in your box.

Threading libraries

You also need threading libraries, which can be downloaded from here. Unpack them to a convenient location, e.g. /usr/local/libraries/:

[local]$ mkdir libraries/

[local]$ tar xzf ethread-lib-jan2012.tar.gz -C libraries/

It is a good idea to create a symbolic link to the threading library. It will allow you to quickly switch to the latest version whenever an update is released:

[local]$ cd libraries/

[local]$ ln -s ethread-lib-jan2012 ethread-lib-latest

Additional software

You will need a whole bunch of additional programs. All these programs are free for academic and non-commercial users; however, some of them do require user registration and/or signing a license agreement. Before downloading and using them, please make sure that you comply with legal and other requirements; we are not responsible for illegal use of third-party software. Moreover, some of these codes are available for packet managers, such as yum (e.g. Red Hat), apt-get (e.g. Ubuntu) or MacPorts (Mac OSX). This is a recommended option and it will save you some time. Additional software is described in the tables below; the detailed installation instructions are provided for each program in the following sub-sections.

Sequence-based tools

COMPASS Protein threading/fold recognition
CS/CSI-BLAST Protein threading/fold recognition
HHpred Protein threading/fold recognition
HMMER Protein threading/fold recognition
NCBI BLAST Sequence alignment
pftools Protein threading/fold recognition
pGenTHREADER Protein threading/fold recognition
PSIPRED Secondary structure prediction
SAM-T2K Protein threading/fold recognition
Sparks/SP3 Protein threading/fold recognition
THREADER Protein threading/fold recognition

Structure-based tools

DFIRE Protein conformation free energy score
DSSP Secondary structure assignment
fr-TM-align Protein structural alignment
Jackal Protein structure modeling
MaxCluster Protein structure comparison and clustering
Modeller Protein structure modeling
Pulchra All-atom reconstruction from the backbone Cα atoms
Stride Secondary structure assignment
TASSER-Lite Protein structure modeling
TM-score Protein structure similarity measure

Machine learning

LIBSVM Support Vector Machines for classification and regression
SVMrank Support Vector Machines for ranking

 

© Michal Brylinski
This website is hosted at the CCT