blast

[AndyOD]

Hi,

Probably a basic question.
I have a file which contains 3 nucleotide sequences.These 3 sequences have been aligned using MUSCLE.

I now want to blast this FILE, as opposed to each gene seperately, against a database of nucleotide sequences, is this possible?

SO basically i want my query to be the file and db to be a massive database of various organisms?



Thanks
Andy

[flxlex]

This is possible both for the web-based blast, as well as for the stand-alone blast. However, you will need to convert your sequences to fasta format.

On the web, just paste the sequences into the query window. The stand-alone (blastall) will take your input file no matter how many sequences it contains.

Happy blasting,

flxlex

[AndyOD]

the reply is very much appreciated.

I think its posibly a little more tricky and involving some kind of profile building?

So I have a file, it has three sequences.I don't want to compare each sequence in that file to a database, but rather I want to see how similar the 3 sequences combined are to a database.

So to do that, I made an alignment of the three sequences.

The next step would be to build a profile of the alignment and then blast the profile against the database, but I can't seem to find any program that builds profiles from nucleotides, prophecy etc. seem to be protein-based?

Confused

[krobison]

It does sound like you wish to build a profile.

First, make sure you understand what you are really trying to do. Why do you believe combining the three sequences will improve sensitivity or specificity or both? What is the underlying pattern you are trying to detect?

There are a host of nucleotide profile building & search tools (e.g. MEME, AlignACE), many of which are focused on looking for short motifs such as transcription factor binding sites. Is this what you expect to find?

Alternatively, you can post-process the BLAST output to ask what sequences were matched by the aligned regions of you queries. If you don't believe there is a sensitivity gain from building a profile, this might give you the required specificity.

AFAIK, there is no mode of BLAST to search nucleotide profiles, but I haven't tracked development of BLAST closely for a while & so could easily be wrong. There are programs to search profiles vs. DNA or can be used to build such profiles; I'm pretty sure HMMER is not strictly protein-based (but again, that hasn't tended to be its main use).

You might also ask whether for what you are looking primary sequence is the most conserved feature. If secondary structure is important, there are other profile construction & search tools at the Eddy lab (HMMER builders) which are more appropriate in that circumstance.