A pattern matching approach to the automatic selection of particles from low-contrast electron micrographs

• DOI: 10.1093/bioinformatics/btt429
• Publication Date: 2013-08-20T01:55:35Z
• Authors: V. Abrishami; A. Zaldívar-Peraza; J. M. de la Rosa-...; J. Vargas; J. Otón; R. Marabini; Y. Shkolnisky; J. M. Carazo; C. O. S. Sorzano
• Keywords: Single-Particle Analysis; Discriminative model; Electron micrographs; Identification; Particle (ecology)

Abstract Motivation: Structural information of macromolecular complexes provides key insights into the way they carry out their biological functions. Achieving high-resolution structural details with electron microscopy requires identification a large number (up to hundreds thousands) single particles from micrographs, which is laborious task if it has be manually done and constitutes hurdle towards high-throughput. Automatic particle selection in micrographs far being settled new more robust algorithms are required reduce false positives negatives. Results: In this article, we introduce an automatic picker that learns user kind he interested in. Particle candidates quickly robustly classified as or non-particles. A discriminative shape-related features well some statistical description image grey intensities used train two support vector machine classifiers. Experimental results demonstrate proposed method: (i) considerably low computational complexity (ii) better comparable previously reported methods at fraction computing time. Availability: The algorithm fully implemented open-source Xmipp package downloadable http://xmipp.cnb.csic.es. Contact: [e-mail] [e-mail] Supplementary Information: data available Bioinformatics online.