Secondary and Tertiary Structure Prediction of Proteins: A Bioinformatic Approach - Complex System Modelling and Control Through Intelligent Soft Computations - page 552

disease) was discovered by docking into a comparative model to dihydrofolate
reductase in L. major, a related parasite (Zuccotto et al. 2001 ). Since the crystal/
NMR structure of various drug targets are not available so far, comparative models
of drug targets could also be used for computational screening of new inhibitors for
Mycobacterium tuberculosis drug target proteins (Gahoi et al. 2013 ).
Comparative modeled structure of cell receptors responsible for binding of
foreign particles and thus causing diseases may also be used to study these inter-
actions and may facilitate in investigating the mechanism. Comparative models can
be also used to predict the antigenic epitopes. Mouse mast cell protease (mMCP) 1,
mMCP-2, mMCP-4, and mMCP-5 models were used to predict immunogenic
epitopes and surface regions that are likely to interact with proteoglycans (Sali et al.
1993 ).
Native PAGE results illustrated the presence of variations in number of isoforms
of superoxide dismutase antioxidative enzymes in different cyanobacterial samples
(Kesheri et al. 2011 ). Comparative modeling may be used to generate antioxidative
enzymes models that may further help in studying the binding of metal cofactors
with the isoforms. Comparative modeling may also be used to study the drug
resistance in many vectors.
Garg et al. ( 2009 ) constructed the comparative model of dihydropteroate syn-
thase protein which illustrated that novel point mutations at two positions may lead
to sulphadoxine drug resistance in Plasmodium falciparum. Compararative models
facilitates molecular replacement in X-ray structure/NMR models which allows
re

nement of a determined structure through the knowledge of already known
structures.The computational prediction of protein structure also serves as an
alternative to produce raw informations that may be validated by wet lab experi-
ments. Following section produces an overview of further developments that may
be made in the

field of protein structure prediction.

8 Future Prospects

Homology modeling and protein threading are becoming more powerful and
important for structure prediction along with the PDB growth and the improvement
of prediction protocols. The error of a template-based model comes from template
selection and sequence-template alignment. So, the identi

cation of the best tem-
plate is still a challenging task in protein structure prediction. However, HMM
based template search algorithms like HHpred has solved this issue to some extent.
Now, another big dilemma is of generation and choosing the correct alignment
between target sequence and template sequence. Still, there is no set benchmark
available for selection of the best alignment between the target and template
sequence.
Model building is also one of the challenging task in structure prediction, in
which a number of times it has been seen that side chains are not added properly in
their proper conformations which mostly need structure re nement. Model