Marker-free image registration of electron tomography tilt-series

• Published in: BMC bioinformatics Vol. 10; no. 1; p. 124
• Main Authors: Sorzano, Carlos Oscar Sanchez, Messaoudi, Cédric, Eibauer, Matthias, Bilbao-Castro, JR, Hegerl, R, Nickell, S, Marco, S, Carazo, JM
• Format: Journal Article
• Language: English
• Published: London BioMed Central 27.04.2009; BioMed Central Ltd; BMC
• Subjects: Algorithms; Bioinformatics; Biomedical and Life Sciences; Computational biology; Computational Biology/Bioinformatics; Computer Appl. in Life Sciences; Computer Simulation; Electron Microscope Tomography - methods; Genetic algorithms; Genetic markers; Image Processing, Computer-Assisted - methods; Life Sciences; Methodology; Methodology Article; Methods; Microarrays; Models, Biological; Phantoms, Imaging; Tomography; France; Spain; Fiducial Marker; Alignment Parameter; Goal Function; Robust Regression; Affine Transformation
• ISSN: 1471-2105; 1471-2105
• DOI: 10.1186/1471-2105-10-124

Background Tilt series are commonly used in electron tomography as a means of collecting three-dimensional information from two-dimensional projections. A common problem encountered is the projection alignment prior to 3D reconstruction. Current alignment techniques usually employ gold particles or image derived markers to correctly align the images. When these markers are not present, correlation between adjacent views is used to align them. However, sequential pairwise correlation is prone to bias and the resulting alignment is not always optimal. Results In this paper we introduce an algorithm to find regions of the tilt series which can be tracked within a subseries of the tilt series. These regions act as landmarks allowing the determination of the alignment parameters. We show our results with synthetic data as well as experimental cryo electron tomography. Conclusion Our algorithm is able to correctly align a single-tilt tomographic series without the help of fiducial markers thanks to the detection of thousands of small image patches that can be tracked over a short number of images in the series.