blast result

Everything on bioinformatics, the science of information technology as applied to biological research.

Moderator: BioTeam

Post Reply
noy k
Garter
Garter
Posts: 8
Joined: Tue Sep 25, 2007 5:00 pm

blast result

Post by noy k » Wed Jul 30, 2008 12:04 pm

how do we know a blast result is significant without relying on the e-value alone?

fcs
Garter
Garter
Posts: 20
Joined: Thu Jul 23, 2009 12:06 am
Location: South San Francisco
Contact:

Post by fcs » Wed Aug 11, 2010 4:46 am

You can also look at the score which gives an indication of how good an alignment is:

http://www.ncbi.nlm.nih.gov/bookshelf/b ... &part=ch16

User avatar
JackBean
Inland Taipan
Inland Taipan
Posts: 5694
Joined: Mon Sep 14, 2009 7:12 pm

Post by JackBean » Wed Aug 11, 2010 8:30 am

the e-value tells you, whether is it significant, why then look to something else? :roll:
score tells you, whether is it good alignment
http://www.biolib.cz/en/main/

Cis or trans? That's what matters.

lishenghui
Garter
Garter
Posts: 4
Joined: Wed Oct 13, 2010 5:56 pm

Post by lishenghui » Thu Oct 14, 2010 5:27 pm

the e-value and the score are entirely positive correlation, and i'm thinking the SCORE is more useful for evaluating a alignment: the e-value 1e-49 and 1e-51 are in the same significant, but the corresponded score 780 and 810 (approximately) are in the difference about 4%........

User avatar
JackBean
Inland Taipan
Inland Taipan
Posts: 5694
Joined: Mon Sep 14, 2009 7:12 pm

Post by JackBean » Thu Oct 21, 2010 6:38 am

as I said above, the score tells you only how good is the alignment. But if you had some short sequence, it may have high score, but be less significant, then worst alignment, yet through longer sequence
http://www.biolib.cz/en/main/

Cis or trans? That's what matters.

Post Reply