The limits of protein secondary structure prediction accuracy from multiple sequence alignment

The expected best residue-by-residue accuracies for secondary structure prediction from multiple protein sequence alignment have been determined by an analysis of known protein structural families. The results show substantial variation is possible among homologous protein structures, and that 100% agreement is unlikely between a consensus prediction and one member of a protein structural family. The study provides the range of agreement to be expected between a perfect secondary structure prediction from a multiple alignment and each protein within the alignment. The results of this study overcome the difficulties inherent in the use of residue-by-residue accuracy for assessing the quality of consensus secondary structure predictions. The accuracies of recent consensus predictions for the annexins, SH2 domains and SH3 domains fall within the expected range for a perfect prediction.