The rockefeller university

by LESLIE CHURCH

In science, seeing the big picture is key. The Rockefeller University Press has taken that literally. Using an online image publishing tool they originally developed in 2008, The Journal of Cell Biology (JCB) has released what it believes is the largest image ever published online — a 281-gigapixel photo of a 1.5 millimeter zebrafish embryo.

The unparalleled level of detail is made possible by an upgrade to the JCB’s DataViewer platform along with newly improved automated image-stitching technology contributed by researchers at the Leiden University Medical Center in the Netherlands. This stitching technique, so-called “virtual nanoscopy”, allows researchers to assemble a mosaic of individual electron microscopy images into one large image. Readers of the journal can zoom in on specific structures of a specimen just as they would zoom in on a house using Google Earth.

The concept of stitching together images to create an ultra-large electron micrograph is not new, but the technology developed by the Dutch researchers allows it to be applied to a much broader range of electron microscopy instruments. Their zebrafish image, made up of more than 26,000 individual images, lets readers view the specimen at the organismal, tissue and subcellular levels, all at high resolution. The JCB worked with the technology company Glencoe Software to develop a platform that could host and present the massive image online.

“The microscope is a vital tool in cell biology, but it’s limited by the small field of view captured in a single image,” says Liz Williams, executive editor of the JCB. “You can get a very high resolution image of a select area of the cell, but if you want to get a broader sense of what’s happening around that area, you lose the resolution of that cellular detail. With virtual nanoscopy, we’re able to integrate high-resolution information across cells and tissues. No other journal provides authors with the technology necessary to share these types of image data.”

The JCB DataViewer has analysis tools that let scientists examine multidimensional microscopy images and interrogate different dimensions — depth, color and time — individually or in parallel. A previous upgrade made the data downloadable in a standardized file format that allows researchers to use their own image analysis software for further examination. The journal also hosts large datasets of quantitative microscopy studies and high throughput image screens.

Authors are not required to submit files to the JCB DataViewer, but since its inception, almost a quarter of the papers published in the journal have also had supplemental image data in the DataViewer. There are currently more than 100,000 images hosted on the site.

“This is the next frontier in scientific publishing,” says Ms. Williams. “The images in print journals are two-dimensional, and the traditional videos in online articles are badly compressed. We provide a new way to share image data at the full resolution at which they were acquired. If you can image it, you should be able to publish it, and now you can.”

Publishing data in its rawest form also ties in with the journal’s reputation for data integrity. Ten years ago it enacted a policy of using Adobe Photoshop to screen images for evidence of manipulation. The Rockefeller University Press has a team that inspects every image in a paper before it’s published.

“The JCB DataViewer provides access to the original data underlying the figures presented in an article. It allows readers to inspect the original data for themselves,” says Mike Rossner, executive director of the press, who pioneered its data integrity policy after he discovered an image in a 2002 manuscript that had been manipulated. “The primary goal for the DataViewer is to enhance data sharing among researchers, possibly leading to new discoveries, but this extra transparency is a welcome byproduct.”