Need Help looking for Motif Search Programs!

[Minkyu]

Hi I'm in need of some help. I need to find a motif with an ambiguous amino acid among sequences that range from 4 to 18 million long. I have a computer with 128 gb of RAM using Windows 7 and I've run motif searches with the programs CLC and Geneious, however, they take too long as they are very processor heavy. I initially tried blast, however, I found that blast does not do well with ambiguous amino acids. Any input would be greatly helpful!

[maasha]

Try scan_for_matches:

http://blog.theseed.org/servers/2010...r-matches.html

[Minkyu]

Thanks! However, after reading the instructions I found that it was CPU intensive not RAM intensive. Do you have any programs that runs primarily on RAM?
Thanks,
Minkyu

[mbblack]

Thanks! However, after reading the instructions I found that it was CPU intensive not RAM intensive. Do you have any programs that runs primarily on RAM?
Thanks,
Minkyu

Well, first all, all programs run in RAM. The issue is that those kinds of searches inherently are cpu intensive. It has nothing really to do with RAM. As long as you have enough RAM to load the search items and query string completely into RAM, then you have maximized your use of RAM and any limitation you experience is the result of computational effort.

The only thing you could do is run the search on a cluster, and partition the search items across multiple nodes (ie. each compute node only searches for a subset of the total sequences in your search set), so no one node has to bear the entire computational load.