Sunday, May 19, 2013

Installing CEGMA...


Installing CEGMA can be a bit of a challenge for most of the users including me! I had installed it earlier, but this time it really gave me a very hard time in RH 6.2. Apart from installing a set of pre-requisitites, installation of genewise is bit of a challenge - especially so because now its status is archival. The INSTALL instruction that comes with the package is not very useful..

Genewise has several pre-requisites:

glib -> Provides low level data structure support for C programming. Can be found here:
http://www.linuxfromscratch.org/blfs/view/svn/general/glib2.html

glib has several pre-requisites such as
1. pkg-config: This package provides tools for passing include and library path so that packages can be built using configure and make utilities.
2. libffi: This library provides high level portable interface for programmers to call any function at run time and can be found at: ftp://sourceware.org/pub/libffi/libffi-3.0.13.tar.gz
This package also requires a patch that can be found at: 
3. Python2.7.5 if it is not already there:  http://www.python.org/ftp/python/2.7.5/Python-2.7.5.tar.xz
4. PCRE library: This contains perl like regular expression libraries:
 http://downloads.sourceforge.net/pcre/pcre-8.32.tar.bz2

All these installation will go smoothly, unless you have some problems with your system. Once done, go
Download wise2.4.1 source from http://wise.sourcearchive.com/downloads/2.4.1/

Do the following:

go to wise-2.4.1/src/HMMer2
then replace 'getline' to 'my_getline' in sqio.c
replace 'isnumber' in src/models/phasemodel.c into 'isdigit'

Then go and check the makefile under each directory under src and relace 'glib-config --libs' to 'pkg-config --libs glib-2.0'  and also glib-config --cflags' to 'pkg-config --cflags glib-2.0' 
using the following command:

find ./ -type f -name "makefile" -exec sed -i.old 's/glib-config --cflags/pkg-config --cflags glib-2.0/g' "{}" +;
and

find ./ -type f -name "makefile" -exec sed -i.old 's/glib-config --libs/pkg-config --libs glib-2.0/g' "{}" +;

Also set path for wiseconfigdir to wisecfg file that is somewhere inside the distribution.

Then do

make all

It works fine for RH6.2


Make sure blast program and Hmmer3 is on your path. Also install geneid in your system as that is also a pre-requisite. Once installed, first set the environment variables for the following:

CEGMA; CEGMATMP; PERL5LIB



PATH=$PATH:$HOME/bin:/usr/adadata/bin:/bin
export WISECONFIGDIR=/usr/adadata/wise2.4.0/wisecfg
export CEGMA=/usr/adadata/cegma_v2.4.010312
export CEGMATMP=/usr/adadata/cegma_v2.4.010312/tmp/
export PERL5LIB=$CEGMA/lib

Now finally Run CEGMA with the default parameters:

bin/cegma --genome sample/sample.dna --protein sample/sample.prot -o sample

References: http://korflab.ucdavis.edu/datasets/cegma/README

2 comments:

Noa z said...

THANK YOU! this id by far the best cegma guide ever!

Sucheta Tripathy PI @ Computational Genomics Group at IICB, Kolkata said...

Good to know that you found it useful...