Publisher:
Universität Dortmund, Sonderforschungsbereich 475 - Komplexitätsreduktion in Multivariaten Datenstrukturen, Dortmund
Abstract:
The secondary structure classification of amino acid sequences can be carried out by a statistical analysis of sequence and structure data using state-space models. Aiming at this classification, a modified filter algorithm programmed in S is applied to data of three proteins. The application leads to correct classifications of two proteins even when using relatively simple estimation methods for the parameters of the state-space models. Furthermore, it has been shown that the assumed initial distribution strongly influences the classification results referring to two proteins.