Modelling Protein Structure from Remote Sequence Similarity: An Approach to Tertiary Structure Prediction
Given some similarity in sequence between two proteins, a molecular model can be constructed for one sequence provided that the tertiary structure of the other is known. This approach, which is often called “modelling by homology”, can yield a reasonably accurate model, providing that the available sequence similarity is clear. With less similarity, useful models can still be constructed and, although these may not be accurate in detail, it is probable that the overall fold will be correct. The robust nature of the approach derives from the stability of the fold of the protein under mutational pressures. This is so conservative that, for two proteins, sequence similarity can be almost undetectable yet the overall fold can remain the same. By implication, if any reliable sequence similarity can be identified at all, then it should be possible to build, at least a rough, model with the correct fold which may well be reasonably accurate in detail in the core.
KeywordsDynamic Program Algorithm Structure Comparison Protein Structure Data Bank Tertiary Structure Prediction Motif Constraint
Unable to display preview. Download preview PDF.
- 1.Taylor, W. R., Orengo, C. A., and Pearl, L. H. (1990). Comparison of predicted and X -ray crystal structures of retroviral proteases. In Ikehara, M., editor, Protein engineering: protein design in basic research, medicine, and industry, pages 21–27. Japan Scientific Press, Tokyo; Springer-Verlag, Berlin. Proceedings of the second international conference on protein engineering.Google Scholar