Protein Structure and Function. Application of Bioinformatics Methods - John Rigden 2014

Comparative Protein Structure Modeling
Conclusion

It has already been proven that comparative modeling is a valuable tool in numerous biological Applications. Its significance among Structure prediction Methods is expected to continue to grow, as the number of experimentally determined protein structures is steadily increasing thanks to Cell/13.html">Protein Structure initiative projects and advances in experimental techniques.

The average sequence identity for structurally related Proteins is typically only 8-9%, with the majority exhibiting less than 15% sequence identity (Rost 1997). Comparative modeling methods are severely limited by this subset of sequences for which sequence similarity to a protein of known structure is detectable. Therefore, it is safe to assume that comparative modeling represents one of the pioneering steps in the field of protein structure recognition and the utilization of structural information. Protein Fold Recognition methods, discussed in Chapter 2, will play a crucial role in extending the capabilities of comparative modeling even to distant homologs and structural analogs.

The key challenges include improving existing methods and developing new ones for the refinement of comparative models through the accurate addition of loops and side chains; optimizing the internal packing of Secondary structure elements; defining scoring Functions capable of comparing Model quality; searching for the optimal combination of known fold fragments; and identifying errors in Spatial Models. Even a minor breakthrough in addressing these issues will have a major impact on comparative modeling, as most relationships between protein structures are too distant to be utilized in comparative modeling approaches. On the other hand, although solving the aforementioned problems is unlikely to dramatically alter the overall accuracy of existing protein models, The Importance of these improvements in yielding functionally reliable spatial models—that is, models that can be confidently used for functional annotation—cannot be overstated.

The aforementioned advances in protein comparative modeling methodologies are essential Prerequisites for the development of novel modeling techniques within the field of "structural Proteomics". The goal of the latter is to assemble the basic Building Blocks of packing models into physiologically more relevant quaternary structures and complexes. This will open up opportunities for modeling the interactions of numerous diverse proteins of known structure.

Acknowledgments. This review is partly based on our earlier publication (Fiser 2004).