blastall blastx filter different hits on one contig

[fisseb4]

Dear all,

I performed a blastx of some thousand contigs of size between 1500 and 5000 bp against uniref-protein database. To speed it up I set the following parameters:

-e 0.00001 -a 8 -b 1 -v 1

so I receive the best hit for each protein. But now I realized that I just got 1 match per contig but on my contigs I expect from NCBI blast 3 or more different hits for example from position on the contig A: 1 - 1000
B 1005 - 2500
C 2600 - 3100 ....... .

Does anybody know what parameter I have to set to get the different hits for A, B, C and so on and not only the 1 best hit of my contig?

Thanks in advance!

[GenoMax]

If you set the b and v option to 1 (http://www.ncbi.nlm.nih.gov/staff/ta...ll_node37.html) then it is only going to show one sequence alignment/description. So to start with increase that number to something bigger. Also the e value cut-off may need to be adjusted downwards to get B and C.

Is there a reason you are still using an older version of blast?

[fisseb4]

I set the b and v option to 1 because I just wanted to get the best result and not something like

UniRef100_U7DV93 Uncharacterized protein n=1 Tax=Pseudomonas flu... 936 0.0
UniRef100_U7DV93 Uncharacterized protein n=1 Tax=Pseudomonas flu... 935 0.0
UniRef100_U7DV93 Uncharacterized protein n=1 Tax=Pseudomonas flu... 933 0.0
UniRef100_U7DV93 Uncharacterized protein n=1 Tax=Pseudomonas flu... 928 0.0
UniRef100_U7DV93 Uncharacterized protein n=1 Tax=Pseudomonas flu... 922 0.0
UniRef100_U7DV93 Uncharacterized protein n=1 Tax=Pseudomonas flu... 919 0.0
UniRef100_U7DV93 Uncharacterized protein n=1 Tax=Pseudomonas flu... 918 0.0
UniRef100_U7DV93 Uncharacterized protein n=1 Tax=Pseudomonas flu... 917 0.0

and so on...
But I did not expect that just 1 hit for the whole contig is shown and not one for the different positions A, B, C which do not overlap. I also tried b and v 30 and like I expected I got some that looked similar to the shown above. And it looked like that every hit was for the same region of my contig

I used blastall because it was installed on our servers. But maybe I should try to install blast+.
I also tried PAUDA and I liked it because it was very fast, but I got a lot of errors and a lot of results were missing.

[GenoMax]

If you are comparing the results from NCBI Blast page then it is certainly using the latest blast+ version. You can find the detailed parameters for NCBI Blast search under the "search summary" link on the blast results page. You can use those to guide your local blast+ search.

[super0925]

If you set the b and v option to 1 (http://www.ncbi.nlm.nih.gov/staff/ta...ll_node37.html) then it is only going to show one sequence alignment/description. So to start with increase that number to something bigger. Also the e value cut-off may need to be adjusted downwards to get B and C.

Is there a reason you are still using an older version of blast?

Hi GenoMax, I have raised a new question about blast, could you please check http://seqanswers.com/forums/showthread.php?p=178574
Thank you!