lufuhao / gicl Goto Github PK

First and important, just to clarify, I am not the author of this program. 



Background:
	GICL/TICL is such a good program to cluster, overlap and extend large set of ESTs. It seems this program is no longer well-maintained, And it's said difficult to setup. And some recent versions of PERL report error. 
	SO, so, so I decided to rewrite the main program to make it run perfectly.
	I believe that would benefit many people, who were or are struggling with EST overlapping problems
	Just to be sure, I did not rewrite those core-algorithm-related executables. So, take it easy to use


	
Description:
	Here is an new repository for GICL, I rewrite some of the perl code to make this program to run with recent perl versions
	GICL is a better and more adjustable program than TGICL. It can be used to cluster, overlap and extend large set of ESTs with or without full-length cDNAs.	



Requirements:
	Linux: csh xorg openbox libxmu-dev libmotif libxp-dev PVM zlib make gcc tar ???
	
	Perl: Getopt::Std; Getopt::Long; Cwd; FindBin; File::Basename; POSIX; Fcntl;



Install:
	
1. First install some requirements
	### But some Linux programs or libraries are needed, like: csh, X11, xmu, motif, Xp
	### Do not install libmotif-dev
	### psx need PVM
	### On my Ubuntu 14.04 x86_64, I installed these by:
	$ sudo apt-get install csh xorg openbox libxmu-dev libmotif libxp-dev zlib1g-dev pvm pvm-dev
	
2. install Perl libs
	

3.	First, comiple NCBI C toolkit v20060507, which is need for the mgblast program. This version of package is included in program subforder, or download it by yourself from NCBI FTP site (ftp://ftp.ncbi.nih.gov/toolbox/ncbi_tools/old/).

	$ cd /yourpath/gicl/programs
	$ tar xvf ncbi_toolkit.20060507.tar.gz

	### Your will see a ncbi folder, and have a look at the ncbi/README or ncbi/make/readme.* files to understand how to compile
	

	###then you need to try 3 times to compile this NCBItoolkit, as it's almost 10 years old. You probably get some warnings when compiling
	$ ./ncbi/make/makedis.csh 2>&1 | tee out.makedis.txt
	### Remember try 3 times until you see this message
			......
			Put the date stamp to the file ../VERSION
			*********************************************************
			*The new binaries are located in ./ncbi/build/ directory*
			*********************************************************
			
	### If you tried 3 times but still got some error, you'd better to check what's wrong based on the error message. GOOGLE it.
	### The files needed should be in /yourpath/gicl/programs/ncbi/include and /yourpath/gicl/programs/ncbi/lib
	
4. Go back to gicl root path which you can find 'install.sh'. Run install.sh to automatically set up all the program

	$ cd /yourpath/gicl/
	$ ./install.sh -i
	
	### OR UNINSTALL all the programs
	
	$ ./install.sh -u

	## IF one of the program failed, the installation will stop. Then you need to go gicl/programs/ folder to check what's wrong based on the make.log and make.err fles
	
5. IF all the programs are successfully compiled, you will see a message to remind you to add /your_full_path/gicl/bin to your PATH, so system knows where to find these executables.

	###try 
	$ /yourpath/gicl/programs/bin/gicl 
	
	### to see if you can see the help message

	That's all, GOOD LUCK.



Usage: 
	$ gicl --query my.fasta -minov 50 --pid 95 --maxovh 40 -mk
	ace2fasta.pl -o contigs.fa *.ace
	cat contigs.fa *.singletons > final.fasta
	
	OR if you have some full length cDNA sequences as references
	
	### Add unique sequence ID prefix to easily filter out these sequences below
	sed 's/^>/>et|/' your.fulllength.cDNA.fa > your.db_et.fa
	cat my.fa your.db_et.fa > xxx.fa
	gicl --query xxx.fa -minov 50 --pid 95 --maxovh 40 -mk -n 2000 -d your.db_et.fa -R et
	### Then filter out those full length sequences
	grep -v '^et|' *.singletons | cdbyank path/to/*.cidx > singletons.fa
    gicl singletons.fa -l 50 -p 95 -v 40 -mk -n 2000
    cat *.ace > zzz.ace
    ace2fasta.pl -o contigs.fa zzz.ace
    cat contigs.fa *.singletons > final.fasta
    


Maintainer:
	Fu-Hao Lu
	Post-Doctoral Scientist in Micheal Bevan laboratory
	Cell and Developmental Department, John Innes Centre
	Norwich NR4 7UH, United Kingdom
	E-mail: Fu-Hao.Lu\@jic.ac.uk