Login

Join for Free!
118908 members


E value confidence

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

Moderator: BioTeam

E value confidence

Postby magicsiew » Mon Dec 13, 2010 7:49 pm

Hi, I have cloned out a gene that without annotation of my particular species in GenBank, when I blast the sequences, able to find the same gene from related species. But the e value are 0.074 and 0.006 respectively. By just analyzing from this data, can I conclude I have successfully clone out the gene? I think these e values are high, but since they are totally different species, so are these e value acceptable? Or is there any parameter of e value for different species?
magicsiew
Coral
Coral
 
Posts: 145
Joined: Sun Jan 24, 2010 3:07 pm

Postby JackBean » Tue Dec 14, 2010 9:42 am

So, you have cloned your gene, sequenced and that sequence used in BLAST, rigth? From this you got 2 entries, with E-values 0.074 and 0.006.

So, what species are you talking about? At least, are they the same genus, order, kingdom(:lol)?
How long was your entry? What was the score and the coverage?
You didn't get any other hits?
http://www.biolib.cz/en/main/

Cis or trans? That's what matters.
User avatar
JackBean
Inland Taipan
Inland Taipan
 
Posts: 5690
Joined: Mon Sep 14, 2009 7:12 pm

Postby magicsiew » Tue Dec 14, 2010 10:11 am

I clone out 2 genes with 0.074 and 0.006 e value respectively. They are totally different species, banana against arabidopsis in Genbank database. My sequences are 850bp and 1kb in length.

The coverage are 2%, very small only, max score is 42.8 and max identical is 98%.

At 1st, I think of doing RACE, but since the coverage is too small, dont really have confidence on it. I try with FASTA and they show other product, rather than the 1 I found using BLAST.
magicsiew
Coral
Coral
 
Posts: 145
Joined: Sun Jan 24, 2010 3:07 pm


Postby JackBean » Mon Dec 20, 2010 2:10 pm

If there's only 2% coverage of 1 kb, that means only 20 similar nts!!! In such a case you cannot take that as significant (that short part may be, but not the whole region)
http://www.biolib.cz/en/main/

Cis or trans? That's what matters.
User avatar
JackBean
Inland Taipan
Inland Taipan
 
Posts: 5690
Joined: Mon Sep 14, 2009 7:12 pm

Postby magicsiew » Mon Dec 20, 2010 5:08 pm

Yes. it cant... Thank you. Hmm... still trying with other methods...
magicsiew
Coral
Coral
 
Posts: 145
Joined: Sun Jan 24, 2010 3:07 pm


Return to Bioinformatics

Who is online

Users browsing this forum: No registered users and 0 guests