Biological Siginificance and estimate probability

G.P.S. Raghava raghava at bioch.ox.ac.uk
Tue Feb 10 11:46:56 EST 1998


Dear All,
I am facing problem in determining correlation between structural
similarity score and in estimate probability score (determined from
PRDF of FASTA by using 1000 suffelles). However, I am getting
very high correlation between structural similaity score and
SD score (Calculated from AMPS on 100 randomization). However,
the correlation in optimised score (PRDF )is giving quite reasonable
correlation. I am not expert may be I am not interprating PRDF output
properly. I am putting PRDF output.

#unshuffled opt score: 291; shuffled score range: 25 - 74
#opt Lambda: 0.18005 K: 0.0157; P(291) = 1.0157e-20
#For 1000 sequences, an opt score >=291 is expected 1.02e-17 times

I am taking '1.02e-17' as estimate probability and '291' as optimised
score from above output. 

I would highly appriciate your help, whether I am calculating estimate
probability wrong or their is other reason, due to which I am not
getting correlation. I expect their should be high correlation between
structure similarity and estimate probability, because it is the measure
of biological significance.

Thanks

Raghava 

-- 
=============================================================
| Dr. G P S Raghava             |            Residance      |
| LMB, Dept. Biochemistry       | 2, Clevedon Court         |
| Oxford University, South Parks| Blenheim Road, Kidlington | 
| Road, Oxford, OX1 3QU, UK     | Oxford, OX5 2JS, Oxford   |
| Email: raghava at bioch.ox.ac.uk | UK                        |
| Email: raghava at biop.ox.ac.uk  | Phone +44-1865-842787     |
| Phone: 275369 Fax: 275182     |                           |
=============================================================




More information about the Mol-evol mailing list