Title :
On the Role of Local Matching for Efficient Semi-supervised Protein Sequence Classification
Author :
Kuksa, Pavel ; Huang, Pai-Hsi ; Pavlovic, Vladimir
Author_Institution :
Dept. of Comput. Sci., Rutgers Univ., Piscataway, NJ
Abstract :
Recent studies in protein sequence analysis have leveraged the power of unlabeled data. For example, the profile and mismatch neighborhood kernels have shown significant improvements over classifiers estimated under the fully supervised setting. In this study, we present a principled and biologically motivated framework that more effectively exploits the unlabeled data by only utilizing regions that are more likely to be biologically relevant for better prediction accuracy. As overly-represented sequences in large uncurated databases may bias kernel estimations that rely on unlabeled data, we also propose a method to remove this bias and improve performance of resulting classifiers.Combined with a computationally efficient sparse family of string kernels, our proposed framework achieves state-of-the-art accuracy in semi-supervised protein remote homology detection on three large unlabeled databases.
Keywords :
bioinformatics; pattern classification; proteins; proteomics; classifier performance; local matching; overly represented sequences; prediction accuracy; protein sequence analysis; semi-supervised protein remote homology detection; semi-supervised protein sequence classification; sparse string kernel family; uncurated databases; unlabeled databases; Accuracy; Bioinformatics; Biology computing; Computer science; Databases; Information resources; Kernel; Labeling; Performance gain; Protein sequence; inexact matching; protein classification; semi-supervised learning; sequence classification; sparse spatial sample kernels; string kernels;
Conference_Titel :
Bioinformatics and Biomedicine, 2008. BIBM '08. IEEE International Conference on
Conference_Location :
Philadelphia, PA
Print_ISBN :
978-0-7695-3452-7
DOI :
10.1109/BIBM.2008.52