There are several steps required to do that, but it is still pretty straightforward, There are probably packages to do that, but I have yet to find one easy to use.

1. Find the genes. (For bacteria I use Prodigal. I used Augustus for Leishmania)

2. Compare the genes. (I usually use blat, which is faster than blast)

3. Filter the comparison to get good hits (you need to set a threshold for identity and a threshold for the length of the alignment vs the lengths of the genes).

4. Group the genes that are considered similar.

5. Parse the file to determine ortholog families.

Does it help?

Hello Artur,

I came across this tool some time ago:

http://cmr.jcvi.org/tigr-scripts/CMR/CmrHomePage.cgi

(under comparative tools)

You can select genome pairs and get the list of proteins that overlap. By exclusion it should be easy to get the ones which are unique to each genome.

I haven't done it myself (not for real, just for fun), but it seems to do what you want.

Abraços,

Chris

I understand you know what to do, just need a few complete genomes to play with. Genomes are all around the Internet here and there in smaller or bigger pockets. To name a few:

http://www.broadinstitute.org/scientific-community/data

http://genome.jgi.doe.gov/

http://www.ensembl.org/info/data/ftp/index.html

Good luck.

Thaks everyone! I got it! Everything is done now...

Best Wishes,

Arthur.