Biomedical Engineering Reference
In-Depth Information
Molecular dynamics involves using Newtonian physics to calculate the force on each atom and move
that atom a distance in a small unit of time. The process is repeated until a pre-determined time limit
is reached. The two major limitations of molecular dynamics in generating tertiary candidates are
round-off error and computation overhead. With longer runs of the program, round-off errors tend to
accumulate. This usually becomes apparent when the total energy of the protein under study begins
to drift. Determining the tertiary structure of a protein based on physical principles alone is
challenging because of the sheer number of possible folds.
Assuming that there are three possible conformations for a given amino acid—a helix, sheet, or
coil—the number of folds is equal to 3
n
, where
n
is the number of amino acids or residues in the
protein. This relationship between the possible number of conformations and amino residue count is
illustrated in
Figure 9-17
. Considering that a molecule with only 100 residues would have 3
100
or 5.2
x 10
47
folds, computationally examining every fold using Newtonian physics would take several
lifetimes on the fastest supercomputer. Generating every possible tertiary structure candidate for a
typical protein, such as Glutamine Synthetase, which has over 5,600 residues, is unlikely without the
invention of fundamentally new form of computing. Currently, generating a 3D structural candidate
with about 50 residues (with 3
50
folds) requires several hours of high-end workstation time. Because
the process may be repeated thousands of times in the creation of a library of candidates, months of
computer time could be involved in the project.
Figure 9-17. Residue Count versus Possible Conformations for Protein
Molecules. The number of possible conformations for protein molecule with
n
residues is 3
n
.
