[Bioperl-l] p-value, e-value

Brian Osborne b_i_osborne@hotmail.com
Mon, 7 Jan 2002 16:09:58 -0500


Jason,

>From the BLAST tutorial at Genbank
(http://www.ncbi.nlm.nih.gov/BLAST/tutorial/Altschul-1.html) :

>This is the P-value associated with the score S. For example, if one
expects to find three HSPs with
>score >= S, the probability of finding at least one is 0.95. The BLAST
programs report E-value rather
>than P-values because it is easier to understand the difference between,
for example, E-value of 5 and 10
>than P-values of 0.993 and 0.99995. However, when E < 0.01, P-values and
E-value are nearly
>identical.

So E and P are not truly synonymous in BLAST, and displaying E appears to be
the way to go since that's what the user sees.

In Sean Eddy's documentation
(ftp://ftp.genetics.wustl.edu/pub/eddy/hmmer/CURRENT/Userguide.pdf) you will
also see him referring to P or P(), distinct from E, so E is probably what
should be displayed since that's what the user sees there as well. I don't
know about FASTA....


Brian O.

----- Original Message -----
From: "Jason Stajich" <jason@cgt.mc.duke.edu>
To: "Bioperl" <bioperl-l@bioperl.org>
Sent: Monday, January 07, 2002 1:40 PM
Subject: [Bioperl-l] p-value, e-value


> I'm implementing the Bio::Search::HSP::GenericHSP and
> Bio::Search::Hit::GenericHit objects, wanted to clarify our use of e-value
> and p-value.  I think we have been interchanging them and considering p
> and alias of evalue.  Is that indeed the correct behavior we should be
> exhibiting?  Do we need to have separate data fields to store these or
> will they be equivalent in all HSP contexts?
>
> In BLAST we have an 'Expect' value in the HSP and a significance value the
> Hit.
>
> In FASTA we have an E() value in the Hit description and an E() value in
> the HSP.
>
> In HMMer we have an E-value in both the HSP and the Hit.
>
> Thanks.
> -jason
>
> --
> Jason Stajich
> Duke University
> jason@cgt.mc.duke.edu
>
> _______________________________________________
> Bioperl-l mailing list
> Bioperl-l@bioperl.org
> http://bioperl.org/mailman/listinfo/bioperl-l
>