Variability, Modelling and Prediction of β-Hairpins with Reference to the Immunoglobulin Fold
In the immunoglobulin (Ig) β-sandwich domain there are two types of connections between sequentially adjacent β-strands. One links adjacent hydrogen-bonded strands — the β-hairpin; the other links strands on opposite sides of the sandwich forming a β-arch. The antibody (Ab) combining site is built from these loop regions, and insertions and deletions found in related sequences commonly occur here. To understand more about the relationship between structure and sequence, we have studied β-hairpins and β-arches in many proteins of known structure, using coordinates taken from the Protein Data Bank. We find that the ‘tight’ hairpins — classified by the lengths and conformations of the loop regions — form distinct families with characteristic sequences. We are using the results of this study to help define template or consensus sequences. These have been incorporated into a supersecondary structure prediction algorithm and are found to improve prediction accuracy. In addition this information can be used in model-building homologous proteins.
KeywordsLoop Region Secondary Structure Prediction Loop Length Improve Prediction Accuracy Tomato Bushy Stunt Virus
Unable to display preview. Download preview PDF.