Hi I would like to know, if anybody of users of this forum use any of the applications mentioned in the topic in real work? If yes, what exactly do you use it for? Or do you know of anybody actually using it?
I spoke to some biologists working in the field but they use just BLAST which is sufficient to them. On the other hand, they told me at support of EBI that they have about 1000 jobs per month.
Thanks.

Hi Azurit.
I have used the standalone SSearch, but probably not for the same application you intend to use it (judging from the other programs you listed). Perhaps you should elaborate on what you want to use these tools for, including the scale of the project. This may help other users better answer your question.

Ryan

Hi Azurit.
I have used the standalone SSearch, but probably not for the same application you intend to use it (judging from the other programs you listed). Perhaps you should elaborate on what you want to use these tools for, including the scale of the project. This may help other users better answer your question.

Ryan

Actually I'm interested in the underlying algorithm (Smith-Waterman) itself more than the application of it in real world and an idea that might improve it came over my mind so I decided to write a paper about it but I would like to know what the real applications are.

In that case I may be able to help. A friend of mine, Gozde, recently completed her thesis on her improvement on a variant of the Smith-Waterman algorithm. You can find her thesis here. I found it to have a very complete introduction which should introduce you to many of the applications and improvements on the dynamic programming algorithms for sequence alignment. Specifically, her application was in the area of small RNA (e.g. microRNA) alignments, both against one-another and against reference genomes. You will also notice that the miRBase web server uses SSearch to identify sequences similar to known miRNAs.


Good luck.

Ryan

I ran shortly through the work (will read it a bit more thoroughly later) but unfortunately there is no example of process where rigorous alignment (SSEARH) is preferred or even it is impossible to use heuristic solution of some kind (i.e. BLAST).

A.

For one example the search of small RNA vs small RNA would be intractable with BLAST, O(m-squared*n-squared) and would not guarantee all matches were found. With Gozde's improved method, you are guaranteed to find a match between any 2 sequences with 3 or fewer differences. Like I said, though, this is not using SSearch, but with Gozde's banded register-level implementation of the algorithm.

Ryan

OK. That's the type of example I search for, thanks. Of course I know that with BLAST you would find everything you would find with Smith-Waterman based algorithms (mainly evolutionary distant protein sequences) but every biologist that I know says that BLAST works just fine, that he/she actually doesn't't need the accuracy they would get with SW solutions.

A.

And I forgot to say that I was also told that SW solutions might be usefull in protein structure prediction but unfortunately that's not very much my field:) But if I understood it right, first when such a prediction is done autamatically, it is inprecise and the prediction is further verified by SW.

I wouldn't know much about that, but back to microRNAs...
If you try both blastn and ssearch on the miRBase website, you will find the latter is much more sensitive when your sequence has more than a few differences than sequences in the miRBase db.

Ryan

Sure. Thanks for this example.

A.