Biology Reference
In-Depth Information
for use with proteins >10-15 aa in length. In contrast, the focused
strategy can start with a limited number of protein variants and still
achieve the desired properties through a guided virtual search among
an astronomical number of protein sequence variants.
At present, the focused strategy is rather limited and complicated.
It cannot be used to build proteins without sufficient knowledge, i.e.
a reliable mathematical model. Although application of the focused
strategy results in identification of the desired protein, it is not true
protein design or engineering but rather a virtual search or protein
sequence space exploration for a protein structure with satisfactory
properties. However, in contrast to the other two strategies discussed
above, this approach builds QSAR knowledge and therefore directly
facilitates development of elementary protein building blocks. Thus,
the focused strategy has two important roles: 1) to build new proteins
with predetermined properties by focusing mathematical modeling
on the protein engineering task, and 2) to build a solid foundation for
the efficient application of the structural strategy by generating ele-
mentary building blocks for protein engineering.
Focused Strategy: Mathematical Modeling
The focused strategy is combinatorial in nature and requires an itera-
tive process that begins with any viral molecule that displays a desired
biological profile and ends with an optimization of the activity profile
for that molecule. The process requires sophisticated computational
techniques to efficiently sample the sequence space for a given appli-
cation. The combinatorial possibilities of this strategy for even simple
systems are explosive and cannot be handled experimentally. For
example, the number of molecules required to interchange 20 amino
acids within the 4-aa peptide is 20 4 , corresponding to 160,000 possi-
ble peptides. The alternative to the experimental approach to protein
optimization is development of a theory, i.e. a mathematical model
that quantitatively relates variations in biological activity to changes in
molecular descriptors (e.g. physicochemical properties of amino
acids) that can easily be obtained for each protein. One mechanism to
reduce the search space is to restrict the types of molecules generated
Search WWH ::




Custom Search