Ambiguous Predictions

The assingnment of secondary structure to either loop, helix or strand was performed by combining the prediction profiles shown in Figure 1 with an analysis of conservation patterns. The strands and helices described in the previous sections show reasonable consistency for one predicted state. However, three regions do not show such consistency and are therefore difficult to assign either to - or - conformation. The region 118-128 starts with a cluster of conserved hydrophobic residues (118-124) indicative of a buried - strand, this is continued by an alternating hydrophobic/hydrophilic pattern suggesting that the buried - strand may extend into a surface strand (125-127). The loop 128-131 conserves proline at 129 and the following region is predicted - (132-135). Type II (proline) turns often occur between consecutive antiparallel - strands, so the conservation of proline lends support to the prediction of - structure on either side of the 132-135 loop. However, this conflicts with the prediction profiles that suggest this region is an - helix.

At 144-150 the prediction profiles suggest - strand, but conserved hydrophobics at 144, 147, 148 and 150 indicate a surface - helix. The tolerance of Gly at 146 may swing the prediction in favour of - rather than - structure.

The variable sequence composition at 330-C-terminus make the region difficult to assign to - or -. The region might be split into two - strands (333-339, 347-353), or possibly a single helix if the gap at 342-344 is closed.