Evaluating statistical significance

J-L Risler risler at cgmvax.cgm.cnrs-gif.fr
Mon May 29 10:33:33 EST 1995

In article <3ps9pi$1ob at mserv1.dl.ac.uk>, cibvm69 at cc.csic.es says...
>Dear netters,
>        I have compared several DNA sequences with Bestfit/Gap. Now, I 
want to 
>get a table of similarities. As these sequences have different lenghts, 
I want 
>to normalize the percent of identities obtained.
>        Is there an alternative to the Q-value that could give me a 
>indicative value similar to the z-score?
>        I have used the RDF2 program for evaluate protein alignments but 
>not sure if this program could be used for DNA sequences. If not, what 
>may I use?


Just use BESTFIT with the option RANDOM: bestfit/rand=100 or bestfit 

This will do exactly the same thing as RDF2, but with bestfit instead of 

Don't get confused: bestfit/random does not calculate the Z-score. It 
just reports the mean "randomized" score and the standard deviation. 
You'll have to get the Z-score by hand.

Of course, this stands for proteins as well as nucleic acids (see also an 
old but interesting paper by Weber et al., (1982) Nucl. Acids Res. 10, 

Good luck,

Jean-Loup RISLER
risler at cgmvax.cgm.cnrs-gif.fr
Centre de Genetique Moleculaire
91198  Gif sur Yvette  France

More information about the Info-gcg mailing list

Send comments to us at biosci-help [At] net.bio.net