Back to MRC-LMB Genomes Homepage

News on geanfammer development. (July 1998)

• Geanfammer can now run PSI-Blast. and parse the output of it for clustering (since Ver. 1.5)

. " GEANFAMMER " refers either to a perl5 program*, a suite of perl5 programs, a perl5 module or a perl5 subroutine library. These are all available by anonymous ftp at cyrah.ebi.ac.uk . It has been developed for the analysis of most of the complete bacterial genomes announced since 1995. It summarises the whole procedure of preparing statistically and biologically more relevant protein (sequence) duplication modules before any more biological analysis like structure and function assignment. With this now anybody can easily analyse the duplication level and types of sequence families in any genome or database.
 

*A very preliminary version of GUI version (Perl/Tk) is also available.(older version)

This is critically important as a lot of protein sequences are multidomain and it can cause serious problems in analysing large amount of sequences automatically if sequences were not broken down to sequence domains.

Geanfammer uses FASTA or SSEARCH which allow the gap in sequence comparison in compared to older BLASTP algorithm which does not. Also, it uses E value instead of Z-score to increase the sensitivity.
 
  DownloadGEANFAMMER single program . A summary of the single program geanfammer.pl is as follows.

The program takes the protein sequences of one or two databases and creates protein families. The protein sequence databases can be a whole genome, part of a genome or any other protein sequence databases in fasta format.

The protein sequence databases are compared to each other (or one database is compared to itself) using one of the two sequence comparison programs of the FASTA package. Using the ouput of the sequence comparison, the proteins are clustered by single linkage. Then GEANFAMMER divides the single linkage clusters which contain unrelated sequences (due to multi-domain proteins) using the DIVCLUS algorithm.

Finally, a sorted cluster file containing the duplication module families is created together with a summary file, which summarizes the distribution of duplication module families.

An example run could be:
 
 
 

prompt> geanfammer.pl YOUR_GENOME.fa

In the distribution, a test fasta format database (geanfammer_test_FASTA_DB.fa) is included, so you can see yourself what it does before trying a bigger real DB. Just type:

prompt> geanfammer.pl geanfammer_test_FASTA_DB.fa

The final result will be "geanfammer_test_FASTA_DB.gclu "
 
 
 

Real Genome TEST!!

We have included the smallest complete Mycoplasma genitalium genome (MG.fa) in the distribution to play with. According to your choice of E value threshold, geanfammer should produce a domain level clustering.

Try:
 

geanfammer.pl MG.fa E=0.2 e=0.2 or
geanfammer.pl MG.fa E=0.01 e=0.01
 

and see what it produces. E=0.2 will produce larger protein families as you are generous in the possible mismatches. E=0.01 can be quite reasonable and we used 0.001 for our genome analysis work to be very strict ( to avoid wrong clusters at the cost of losing distant but true members). The search part of the program will take the most time. It will produce a subdirectory called MG in which the results of search will be stored. Final results will be made in the present directory. So, it is a good idea to make a new directory for the test and run geanfammer inside it.
 

A documentation of the single programs follows here, although details on usage can be found in the headers of all programs:

• The search scripts are do_fasta_sequence_search.pl, do_ssearch_sequence_search.pl or do_sequence_search.pl. They take as arguments a query database, a target database and the path and name of the search program (either FASTA or SSEARCH respectively). do_sequence_search.pl do self self search as default.

• Next, all the ssearch output format files (.msso files) generated by the search script have to be converted to msp (matching sequence pair) format files. This is done using sso_to_msp.pl. sso_to_msp.pl has to be run in the directories with the sso files as follows : "sso_to_msp.pl *.sso (ie all sso files) e". The e option means that 1 msp file is created for each sso file, rather than a single msp file for all sso files. The header of the program can be read for further details of options.

• Next, the msp files can be read by msp_single_link.pl to create a file with single linkage clusters of the data at a given expectation value (E-value) threshhold. msp_single_link.pl has a single argument, the E-value threshhold. A single file "single_linkage.sclu" is created by this program as default unless any name is given.

• msp files for each single_linkage cluster must then be created by sso_to_msp.pl. Here, sso_to_msp.pl must be run as follows: sso_to_msp.pl single_linkage.sclu.

• The resulting msp files can then be put into DIVCLUS, which has its own page. DIVCLUS makes .clu files containing the duplication module families. There are two important options: e, the e-value threshhold and f, the minimum percentage overlap region. Usually, e should be the same as the single linkage threshhold. f should be 3 (=67%) or higher (for clusters larger than 100 sequences). DIVCLUS is run in the form: "divclus.pl e=0.001 f=3 *_cluster.msp".

• A program which creates a single file, called "sorted_cluster_file.gclu", of all the families is create_sorted_cluster.pl. It is simply run in the directory with all the clu files produced by DIVCLUS.

• make_clustering_summary.pl provides a summary of the families in the "sorted_cluster_file.gclu" in "sorted_cluster_file.summary". This also inserts the summary to "sorted_cluster_file.gclu".

. You can also download geanfammer from CPAN site. However, it might not be as updated as above ftp routes.
 
 
 

• CPAN (comprehensive perl archive net) HOME site (most uptodate, can be slow),
• CPAN Hensa UK mirror

. Related WWW sites

GeneQuiz

Domainer (Prodom), Prodom HOME, Prodom DB search
 
 
 

. Warranty
 
 
 

We are the programmers who made this, so we will do our best to tackle any problems you have while using the program(s). However, there is no legal guarantee on the possible malfunction of any part of the package.

Bug report -->  sat@mrc-lmb.cam.ac.uk , jong@mrc-lmb.cam.ac.uk

(C) Copyright. 1995.

Free for academic research and educational purposes for non-profit making purposes. The copyright rule for Perl itself applies to the program(s). For commercial use and collaboration please contact the authors.