A Self-Consistent Method for the Analysis of Protein Secondary Structure from Circular Dichroism

Abstract

A self-consistent procedure for estimating the secondary structure content from circular dichroism spectra of proteins is presented. In this method the spectrum of the protein to be analyzed is included in the basis set and an initial guess is made for the unknown structure as a first approximation. The resulting matrix equation is solved using the singular value decomposition algorithm and the initial guess is replaced by the solution. The process is repeated until self-consistency is attained. The best features of the variable selection and the locally linearized methods are incorporated in this procedure. We have applied this method to examine the inconsistencies in the CD data, to compare the predictions with different ranges and resolutions of the CD data, and to compare different assignments of secondary structures from X-ray structure analyses in the context of secondary structure predictions. The results are compared using the root mean square differences and correlation coefficients. The results obtained are as good as or better than the previous analyses. For most of the proteins considered the self-consistent solutions obtained with different initial guesses were similar. We find the Kabsch and Sander protein crystal structure analysis to be most suitable for our prediction method.