%0 Journal Article
%T Local protein structure prediction using discriminative models
%A Oliver Sander
%A Ingolf Sommer
%A Thomas Lengauer
%J BMC Bioinformatics
%D 2006
%I BioMed Central
%R 10.1186/1471-2105-7-14
%X We developed a local structure prediction method to be integrated into either fold recognition or new fold prediction methods. For each local sequence window of a protein sequence the method predicts probability estimates for the sequence to attain particular local structures from a set of predefined local structure candidates.The first step is to define a set of local structure representatives based on clustering recurrent local structures. In the second step a discriminative model is trained to predict the local structure representative given local sequence information.The step of clustering local structures yields an average RMSD quantization error of 1.19 £¿ for 27 structural representatives (for a fragment length of 7 residues). In the prediction step the area under the ROC curve for detection of the 27 classes ranges from 0.68 to 0.88.The described method yields probability estimates for local protein structure candidates, giving signals for all kinds of local structure. These local structure predictions can be incorporated either into fold recognition algorithms to improve alignment quality and the overall prediction accuracy or into new fold prediction methods.In recent years progress has been made in protein structure prediction by incorporating information on local protein structure. David Baker's group has successfully used local fragment predictions [1-4] in conjunction with a fragment assembly procedure to substantially improve new fold predictions [5-7]. Also for fold recognition and remote homology detection methods the integration of local fragment predictions led to improved results [8].Methods for analyzing fragments focus on sequence or structure or both. We are looking for fragments that occur in several proteins, that are sufficiently similar in structure, and that exhibit enough sequence similarity to be detectable by discriminative methods.Specifically, we address the question: Given a local sequence fragment, how much can we learn about the lo
%U http://www.biomedcentral.com/1471-2105/7/14