By Liat Clark Courtesy of WIRED.CO.UK

A team of molecular biologists has published a paper revealing how its "virtual nanoscopy" method creates detailed, high resolution images of cellular structures by "stitching" together thousands of electron microscope photos.

Electron microscopes can magnify an image by up to 10 million times using beams of electrons. However, the tool can only be used to either capture a single, detailed image of part of a cell or, at a lower resolution, a less-detailed overview of the cell. There was no way, until now, to relate the one to the other and give a contextual and detailed overview of the entire cell structure at once.

In the paper Virtual nanoscopy: Generation of ultra-large high resolution electron microscopy maps, Leiden University Medical Centre molecular biologist Frank Faas and his team explain how they have enabled "unbiased high resolution data access while maintaining the lower resolution overview of the cellular context". It's a little like Google Earth's zoom tool, but for cell biologists.

The team joined together 26,434 detailed photos of a 1.5 millimetre-long zebrafish embryo taken using an FEI Eagle CCD camera linked up to an electron microscope. The total data of all the photos that made up the full embryo picture amounted to 281 gigapixels, with a resolution of 16 million pixels per inch -- to visualise just what incredible detail that resolution offers, compare that pixel count to the 220 pixel-per-inch resolution boasted by Apple's new Retina display Macbook Pro. Seems pretty paltry now, doesn't it. The final image slide was created using a program specifically developed for the task -- MyStitch extracts metadata from transmission electron microscopy (TEM) images and uses this to pair the images, noting any overlaps and adjusting the joins appropriately. Images have been stitched together in a similar manner in the past, using tools such as Adobe Photoshop, but only on a much smaller scale. Generating a full image of a cellular structure calls for huge amounts of time-consuming data collection, so the team achieved its feat by automating some of the processes, from the data-stitching down to the photo-taking.

The final image has been published at the upgraded JCB DataViewer site where interested parties can play around with the zoom function and take in not only the embryo, but images of mouse glomeruli, human dendritic cells and mouse embryonic fibroblasts.

Medical professionals often zoom in on the part of a cell they believe to be the main area of focus, or the root of a problem. By presenting a comprehensive image, the new "virtual nanoscopy" method ensures they get the whole picture and do not miss or inadvertently skim over important elements.

"Virtual nanoscopy does not suffer from sparse or possibly biased selection of regions of interest for high resolution imaging," states the paper. "[It] provides an objective and representative approach to record, communicate, and share data of large areas of biological specimens at nanometre resolution."

And in a related story from newswise (Rockefeller University Press):

Virtual Nanoscopy: Like “Google Earth” for Cell Biologists

Since the early days of cell biology, electron microscopy has revealed cellular structures in exquisite detail. The technique has always been limited, however, by the fact that it can only capture a tiny portion of the cell in a single image at high resolution, making it difficult for researchers to relate the structures they see to the cell as a whole, let alone to the tissue or organ in which the cell is located. Viewing samples at lower resolution, on the other hand, can reveal the larger picture of a cell or tissue, but researchers then lose the benefit of seeing fine details.

A team of scientists from Leiden University Medical Center in the Netherlands has addressed this problem by developing new tools for stitching together thousands of electron microscopy images into single, high-resolution images of biological tissues—a “Google Earth” for cell biologists—which can be explored using the newly enhanced JCB DataViewer.

Faas et al. describe their recent advances to a technique called “virtual nanoscopy” in the August 6th issue of JCB. The researchers were able to stitch together over 26,000 individual images to generate an almost complete electron micrograph of a zebrafish embryo encompassing 281 gigapixels in total at a resolution of 16 million pixels per inch. Using the JCB DataViewer, anyone can navigate the zebrafish image from the level of the whole, 1.5 millimeter-long embryo down to subcellular structures.

The ability to integrate information across cells and tissues will provide researchers with exceptional opportunities for future discoveries. But the image’s large size and complexity meant that providing access to Faas et al.’s data necessitated a major upgrade to the JCB DataViewer, a browser-based image hosting platform originally launched in 2008 to promote the sharing of original data associated with JCB publications.

“If you can image it, you should be able to publish it,” says JCB Executive Editor Liz Williams. As a journal, “JCB remains committed to developing cutting-edge tools for the presentation of the data that drive progress in the field of cell biology.”
# # #
About The Journal of Cell Biology 
Founded in 1955, The Journal of Cell Biology (JCB) is published by The Rockefeller University Press. All editorial decisions on manuscripts submitted are made by active scientists in conjunction with our in-house scientific editors. JCB content is posted to PubMed Central, where it is available to the public for free six months after publication. Authors retain copyright of their published works and third parties may reuse the content for non-commercial purposes under a creative commons license. For more information, please visit http://www.jcb.org.

About the JCB DataViewer
The JCB DataViewer is an image hosting and presentation platform for original image data associated with articles published in JCB. Developed in a collaboration between Glencoe Software, Inc. (www.glencoesoftware.com) and the Rockefeller University Press (www.rupress.org), the JCB DataViewer was the first browser-based viewer for multidimensional microscopy image data. It is based on open source software built by the Open Microscopy Environment (OME; http://openmicroscopy.org).

Faas, F.G.A., et al. 2012. J. Cell Biol. doi:10.1083/jcb.201201140
Williams, E.H., et al. 2012. J. Cell Biol. doi:10.1083/jcb.201207117