Unsupervised Integration of Multiple Protein Disorder Predictors: The Method and Evaluation on CASP7, CASP8 and CASP9 Data

Proteome Science

Proceedings

Zoran Obradovic¹  Ping Zhang¹ 

[1] Center for Data Analytics and Biomedical Informatics, Temple University, 19122, Philadelphia, PA, USA;

关键词: Prediction Performance;    Area Under This Curve;    Probabilistic Label;    True Label;    Component Predictor;   

DOI  :  10.1186/1477-5956-9-S1-S12

来源: Springer

【 摘 要 】

BackgroundStudies of intrinsically disordered proteins that lack a stable tertiary structure but still have important biological functions critically rely on computational methods that predict this property based on sequence information. Although a number of fairly successful models for prediction of protein disorder have been developed over the last decade, the quality of their predictions is limited by available cases of confirmed disorders.ResultsTo more reliably estimate protein disorder from protein sequences, an iterative algorithm is proposed that integrates predictions of multiple disorder models without relying on any protein sequences with confirmed disorder annotation. The iterative method alternately provides the maximum a posterior (MAP) estimation of disorder prediction and the maximum-likelihood (ML) estimation of quality of multiple disorder predictors. Experiments on data used at CASP7, CASP8, and CASP9 have shown the effectiveness of the proposed algorithm.ConclusionsThe proposed algorithm can potentially be used to predict protein disorder and provide helpful suggestions on choosing suitable disorder predictors for unknown protein sequences.

【 授权许可】

CC BY   
© Zhang and Obradovic; licensee BioMed Central Ltd. 2011