IMPROVING PROTEIN SECONDARY STRUCTURE PREDICTION WITH ALIGNED HOMOLOGOUS SEQUENCES

Most recent protein secondary structure prediction methods use sequenc e alignments to improve the prediction quality. We investigate the rel ationship between the location of secondary structural elements, gaps, and variable residue positions in multiple sequence alignments. We fu rther investigate how these relationships compare with those found in structurally aligned protein families. We show how such associations m ay be used to improve the quality of prediction of the secondary struc ture elements, using the Quadratic-Logistic method with profiles. Furt hermore, we analyze the extent to which the number of homologous seque nces influences the quality of prediction. The analysis of variable re sidue positions shows that surprisingly, helical regions exhibit great er variability than do coil regions, which are generally thought to be the most common secondary structure elements in loops. However, the c orrelation between variability and the presence of helices does not si gnificantly improve prediction quality. Gaps are a distinct signal for coil regions. Increasing the coil propensity for those residues occur ring in gap regions enhances the overall prediction quality. Predictio n accuracy increases initially with the number of homologues, but chan ges negligibly as the number of homologues exceeds about 14. The align ment quality affects the prediction more than other factors, hence a c areful selection and alignment of even a small number of homologues ca n lead to significant improvements in prediction accuracy.