COMP: statistical methods for peptide amino acid composition

[Jurg Ott <ott@linkage.cpmc.columbia.edu>]

Date: Fri, 8 Mar 1996 19:13:32 +0300
From: Vladimir Morozov <morozov@ENZYME.CHEM.MSU.SU>

        I apply statistic methods for analysis of dependence of
functional property (activity, stability, ED50, Ks e.s) on amino
acid composition in a peptide set.  It is especially promising for
identify "key" residues (mostly influencing on properties) in
random peptide libraries and  protein families divided by
functional, evolution,  or other criteria.
        I would like to collaborate with groups which need the such
analysis their data.
        I use the methods incorporated into the ProAnalyst package:
- discriminant and multiple linear regression analysis of
relationships between functional property and physic-chemical
parameters of an amino acid site.
- analysis of variations (ANOVA) and some other.

Vladimir Morozov

Chemical Enzymology Department,  Faculty of Chemistry,
Moscow State University, Russia
Fax:007-095-939-27-42 or 007-095-939-35-89
E-mail: morozov@enzyme.chem.msu.su