Biology Reference
In-Depth Information
Peptide and Protein Design
Empirical methods for building predictive models of the relationships
between molecular structure and functions of molecular candidates are
becoming increasingly important for vaccine discovery and develop-
ment. Automation of chemical synthesis, tools for designing molecule
libraries and high-throughput technology for biological screening
experiments robustly and quickly enable target candidates to be dis-
covered more effectively. As vaccine and diagnostic candidates progress
down the development pipeline, the ability to predict their immuno-
logical, antigenic, pharmacokinetic, and toxicological properties is
becoming increasingly important in reducing the number of expensive
late-development failures. As a result, the application of ANN methods
in this area of research has gained popularity.
217-226
The major ration-
ale for the use of ANN methods for the prediction of antigenic/
immunogenic properties of viral molecules is that determining for
which structural parameters the ANN best predicts may lead to the
identification of relevant features responsible for the observed activities.
Extracting such knowledge from primary structure is necessary for the
successful design and engineering of molecule candidates with desired
properties. In addition, QSARs are usually complex and nonlinear, and
the ANN is often chosen because the system tends to provide more
flexibility and adaptability for modeling non-linear relationships and is
more noise-tolerant compared to other methods.
217
Schneider
et al
. (1998(b); 12179-12184) proposed a systematic,
focused strategy for peptide design through a combination of rational
and evolutionary approaches.
227
They described the use of evolution-
ary strategies as a cyclic-variation selection of peptide candidates fol-
lowed by ANN modeling of quantitative structure-activity
relationships. They reported successful application of an approach to
the
de novo
design of a peptide that fully prevents the positive
chronotropic effect of anti-b
1
-adrenoreceptor autoantibodies from
the serum of patients with idiopathic dilated cardiomyopathy (DCM).
Based on previous reports showing that short synthetic peptides
encompassing the natural epitopes were able to neutralize the
chronotropic effect of the autoantibodies, their idea was to test
