Figure 5

Methodology for comparative modelling. Target sequences related to the sequences with conformations determined by experiment were candidates for comparative modelling. Generally alignments were obtained from the various servers available as part of the CAFASP meta-server [29]. Initial models were then be constructed and structure-based alignments were used in an iterative manner to refine alignments manually. Non-conserved side chains and main chains were constructed using a graph-theoretic approach with sampling provided by exhaustive and database searches. The final conformations were minimised by ENCAD [36].