Grasping molecular structures through publication-integrated 3D models Pravin Kumar 1, * , Alexander Ziegler 2, * , Julian Ziegler 1, 3 , Barbara Uchanska-Ziegler 1 and Andreas Ziegler 1 1 Institut fu ¨ r Immungenetik, Charite ´ –Universita ¨ tsmedizin Berlin, Campus Benjamin Franklin, Freie Universita ¨ t Berlin, Thielallee 73, 14195 Berlin, Germany 2 Institut fu ¨ r Biologie, Freie Universita ¨ t Berlin, Ko ¨ nigin-Luise-Strasse 1-3, 14195 Berlin, Germany 3 Current address: Meister-Francke-Strasse 8, 22309 Hamburg, Germany Although the need for communicating 3D data using simple and intuitive means extends to disciplines as diverse as biology, engineering sciences and the visual arts, none of the currently available molecular-visualiza- tion programs depicting potentially highly complex structures are compatible with the portable document format (PDF), the current gold standard of electronic publishing. Therefore, it is highly desirable for authors to be able to provide their readers with a basic 3D display of structures that can be accessed without the need for specialized visualization software. Here, we describe how an interactive 3D model of a molecular complex can be embedded directly into a PDF, thus providing readers with important and educational visual information that would otherwise be more difficult to disseminate. Benefits of portable document format (PDF)-integrated 3D models How multidimensional research data are presented in a scientific publication will often be crucial for an in-depth understanding. This is not only the case in structural biology where molecular features have been accessible in print format for many years through static and ‘stereo’ images [1–5], but is also the case in many other disciplines, such as chemistry, physics or medicine. However, complete access to a 3D model of a molecule that enables direct interaction with all its components, requires not only ad- equate hardware, but also the skill to operate a molecular- visualization tool appropriately, including the ability to effectively obtain and extract the precise information that is desired. Because these programs are often complicated to handle, there is an urgent need for a simple, effective and interactive representation of 3D structures that would enable both non-specialist and specialist readers of scien- tific publications to gain access to the most important aspects of the information contained within such struc- tures. This requirement has already been recognized but there is as yet no consensus as to which presentation format is best suited [6,7]. Some examples of attempts to address this issue include internet browser plug-ins, such as Protein Explorer [4] and Jmol [7,8], which provide the possibility of integrating interactive 3D models into scientific and educational web- pages, and the iSee platform [6], which offers a unified access system to a variety of components of structural genomics data from genes to complete structures. How- ever, all these technologies require installation of pro- grams or plug-ins by the user before the relevant information can be accessed and, crucially, the information is only accessible when online. The alternative or comp- lementary mechanism is to provide a pre-rendered film of the molecule in a downloadable video format. Although improving the ‘viewability’ for the reader of the 3D infor- mation relative to that of static 2D images, this prevents any direct interaction with the structure, thus limiting its potential educational benefit and restricting the reader to scenes of the molecule dictated by the author. An additional problem for both of these approaches is that the information has to be supplied as a supplement to the original article rather than as a part of it, thus proving an impediment to the reader’s access. Significantly, none of these technologies are compatible with the Adobe Systems’ PDF, the universally accepted standard of electronic pub- lications. A recent article [9] exemplifies these problems: although a 3D viewer (‘FirstGlance in Jmol’) with several different viewing options can be activated through a supplementary attachment, all attributes are only avail- able online in a standardized manner through Jmol, pre- cluding any integration into a PDF file, the format by which the article is distributed. To simplify communication between the author and the reader, the aim must therefore be to depict structural data with the maximum of information about the molecules described but integrated within the electronic publication itself. Here, we show how an interactive, highly informa- tive 3D model that fulfils these requirements can now be assembled and integrated into PDF documents. In addition to the typical 2D representations of static figures, this leads to enhanced, directly accessible 3D information in the content of publications in which structural data are described, thus combining the benefits of all other approaches to the dissemination of 3D structural infor- mation into a single, standard format file. The HLA-B*2705–pVIPR complex To demonstrate how a 3D model can be created and integrated into a PDF file, we chose the structure [Protein Opinion Corresponding author: Ziegler, A. (andreas.ziegler@charite.de). * Joint first authors. TIBS-596; No of Pages 5 0968-0004/$ – see front matter ß 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.tibs.2008.06.004 Available online xxxxxx 1