A new set of descriptors called molecular electro - negativity edge vector
(MEEV) were developed from the primary structures of peptides, in order to
describe the peptide structure, based on the distance between atoms and the
electro - negativity of each atom. Several models of quantitative structur
e activity relationship are proposed for 58 dipeptides (angiotensin convert
ing enzyme inhibitors) by the multiple linear regression method. In order t
o explain the ability to characterize the molecular structure of polypeptid
es, a model investigation was carried out on the QSAR of pentapeptides. The
results showed that MEEV exhibited both excellent structural selectivity a
nd good estimation of activity. MEEV is easy to use and generally applicabl
e. There fore, it should be very useful in structural characterization and
activity prediction of biological molecules.