SEQanswers

SEQanswers (http://seqanswers.com/forums/index.php)
-   Bioinformatics (http://seqanswers.com/forums/forumdisplay.php?f=18)
-   -   Interesting Behavior: The megablast standalone behaves as blastn.. Who knows why??? (http://seqanswers.com/forums/showthread.php?t=19796)

pbrand 05-04-2012 07:45 AM

Interesting Behavior: The megablast standalone behaves as blastn.. Who knows why???
 
Hi all,
I am using the blastn standalone with the MEGABLAST algorithm to find contaminations in my illumina Hymenopteran database prior to assembly.

I observed the interesting instance, that when I use the nuccore db as database to blast against, the found best blast hits aren't the same as when I use a custom Hymenoptera database (all hymenopteran sequences of ncbi) without changing any other settings.

Now you think: "Clear thing, it's because of the databases! You cannot find anything that is not there"

BUT:

When I test reads with hits on both approaches on the ncbi-web-blast, I get the same behavior as described above when I search one time using the MEGABLAST algorithm (the same hit as against nuccore) and the other time using the blastn algorithm (the same hit as against hymenoptera-DB) WITHOUT changing the database!!!!!!!

(just to point it out: that means, that the behavior is not related to the different databases but to the algorithm!)

You can try it at the web blast if you want:

GAAAAAAAGACAGTTGCGTATCATGAGGCTGGCCATGCAGTAGCAGGCTGGTTTTTACAATACGCAGATCCACTTTTAAAGGTCTCTATAATCCCACGTGG

bbh with megablast: NM_001079291.1 Xenopus e-15
bbh with blastn: XM_003693554.1 Apis e-28

bbh standalone megablast nuccore: NM_001079291.1 Xenopus e-15
bbh standalone megablast hym-DB: XM_003693554.1 Apis e-28

That is totally strange, isn't it?

Why does my standalone megablast find Apis just when I use the custom hymenopt. DB, not in the nuccore? Does it act as dc-megablast or blastn by itself in special circumstances?

Does anyone know if the standalone blastn Megablast changes any settings like word size on it's own? Can anyone explain this strange behavior?

I would be happy to hear any of your expectations.

Cheers,
Phil


All times are GMT -8. The time now is 10:00 PM.

Powered by vBulletin® Version 3.8.9
Copyright ©2000 - 2020, vBulletin Solutions, Inc.