[Bioperl-l] Megablast Ouput

Waibhav Tembe tembe at bioanalysis.org
Tue May 3 12:02:44 EDT 2005


Thanks for the quick reply. Another question:

If I don't care about the statistics of the the hits, but only need the 
local alignments and no. of gaps/mismatches, is using megablast and 
blast with the following parameters equivalent?

blastall -p blastn -d DB -F f -q -1 -r 1 -G 2 -E 1 -W 7 -I T -e 10000000 
-i Input -o Output

megablast -d DB -F f -q -1 -r 1 -G 2 -E 1 -W 28 -I T -m 0 -D 2 -n T -R 
T  -i Input -o Output

1. I tried to set E value large in BLAST to get all possible hits.
2. Since megablast computes a hash for every 4 base pairs, I selected 
W=28 which (I think) should match W=7 for blastn

Any suggestions/comments/criticism welcome.

I have to blast more than 100 short sequences of length approx 50bp and 
need to get local alignment scores using a modest linux box. I observed 
that megablast runs very fast as compared to BLAST. So I am curious to 
know if I will miss any hits using megablast instead of blast. As I 
mentioned, I don't care about the statistics.

Thanks.


Jason Stajich wrote:

> Bio::SearchIO::blast will parse MEGABLAST output the -D 0 and -D 2 
> formats.
>
> -jason
>
> On May 3, 2005, at 11:40 AM, Waibhav Tembe wrote:
>
>> Hello List,
>>
>> I was wondering if BioPerl parse routines from BLAST output could be 
>> used for megablast output as well (assuming that the display format 
>> for megablast output has been set to BLAST like output).
>>
>> Secondly, if the output display format (-m parameter) for blastall is 
>> changed, is it still possible to use BioPerl blast output parsers to 
>> extract hits/HSPs/Evalue etc?
>>
>> Thanks.
>>
>> Regards,
>>
>> -waibhav
>>
>> _______________________________________________
>> Bioperl-l mailing list
>> Bioperl-l at portal.open-bio.org
>> http://portal.open-bio.org/mailman/listinfo/bioperl-l
>>
> -- 
> Jason Stajich
> jason.stajich at duke.edu
> http://www.duke.edu/~jes12/
>
>





More information about the Bioperl-l mailing list