One of the coolest things I get to do at MindTouch is learning about our customers. We have some of the most amazing clients who are using Deki to manage and share ideas; seamlessly communicate no matter the distance, update their network, the list goes on. The best part about getting to know my customers is to hear their stories; stories full of hope, vision and passion - about the challenges they face, the successes they have had, and the greater opportunities they seek. It’s these stories that give us the encouragement to do even better for our customers – the ones who endlessly inspire us here at MindTouch.

One such customer is the Bioinformatics Core of the Joint Center for Structural Genomics, located right here in San Diego, and housed at the University of California, San Diego and the Burnham Institute for Medical Research. The JCSG is one of the four large-scale structural genomics centers in the United States funded by the National Institute of General Medical Sciences as part of their Protein Structure Initiative. The goal of PSI centers is to dramatically reduce the costs and time to determine the three-dimensional shapes of proteins at an atomic level and eventually, to make the structures of most proteins easily obtainable from knowledge of their corresponding DNA sequences. In a general sense, DNA sustains a set of instructions and provides them to direct the processes of life, but the proteins are the central machinery for those processes. Moreover, proteins share some common features in their architecture, such as, for example, segments looking like the coiled coils of telephone cords, and the organization of these features are critical for how the protein machinery works, what it does and when and where. A very small change in the structure of a protein can have a very profound impact, including giving rise to disease or death. I have discovered there is a lot of passion to be found in scientists who study proteins.

Scientists at the JCSG have developed a wiki-based platform, The Open Protein Structure Annotation Network (TOPSAN), as a new model for creating and disseminating knowledge about protein structures emerging from structural genomics centers. Their wiki on protein structures draws information from multiple databases and content repositories and provides for a collaborative environment for researchers from all over the world to contribute their expertise in figuring out the functions of these proteins

Sounds pretty amazing, huh? Makes you marvel at how a passion for proteins when combined with the collaborative power of wikis can enable new science and serve society better. So naturally, I jumped at the opportunity to speak with the founder of the project, Srikrishna Subramanian. Here’s what he had to say about TOPSAN and how Deki has facilitated and helped attain their goals.

Tell me more about TOPSAN and its goals.

Structural genomics efforts in the United States have determined the three-dimensional structures of more than 3000 proteins, most of which are novel and previously uncharacterized, as part of their mission to systematically sample the range and diversity of protein structures. Innovative methods have been established to yield a very large output of new protein structures. These structures are known with great precision, as specific in quality or better than how much has been ascertained about proteins studied by traditional methods. At the same time, the novelty of the proteins and the rapid pace at which these protein structures are determined precludes prior knowledge and the potential for extensive local efforts in establishing deep insight into the function of these proteins. When additional information about the implications of the chemical, physical and biological properties of a protein exists, this information is recorded with the structure and the process is termed annotation. For the vast number of novel protein structures determined through structural genomics, little therefore is known beyond the structure itself about these proteins, and consequently, many of them often remain unnoticed by other researchers, even those directly working on the same or related proteins. In order to address this problem, we established TOPSAN in July 2006 as a collaborative platform, where researchers from around the world can participate, discover and create “new knowledge” about our protein structures. We sustain traditional academic approaches for quality control and for authorship. Overall, TOPSAN attempts to engage the structural biology experimentalists and the world of biology researchers in an active dialogue to characterize these novel protein structures. The potential even exists for TOPSAN users to establish enough new knowledge to lead to extensive collaborations with the researchers in structural genomics that determined the structure.

How has MindTouch Deki helped with your efforts?

We initially evaluated a proprietary wiki, JotSpot (currently owned by Google), which gave us considerable experience with the concepts and ideas on what we needed to do for the biological science community to initiate and sustain a dialogue on novel protein structures, but in the process we also discovered the limitations and risks of proprietary software (even if it is free!). After Google’s acquisition of JotSpot, it was eventually re-packaged (as Google Sites) and many of the core features that TOPSAN required were not relevant for Google and thus, no longer supported. An open source platform enables stability in such a situation and extends us the flexibility to develop and tailor the platform to suite our specific requirements. The Deki platform developed under the GNU General Public License, together with the extensive support we receive from engineers at MindTouch, provides for a more reliable and sustainable path for us to fully establish TOPSAN as a flexible, scalable and extensible knowledgebase environment that can be readily replicated, and extended by others.

What’s your favorite thing about using MindTouch Deki for TOPSAN?

Several things, really… From a developer’s viewpoint, the best thing about Deki is that its code is open source (no surprise there!). Other benefits include the relative ease with which data from other sources can be streamed into TOPSAN (Mashups) and similarly, the ease of backend access to Deki via the API that allows us to not only automatically start new wiki pages but also allows other resources to access content from TOPSAN. However, from a user’s perspective, I would say the favorite thing is the ease of use of Deki for beginners without sacrificing functionality. Users of TOPSAN are researchers whose primary expertise is in biology; many of them have not utilized wikis in their day-to-day research activities but are nevertheless excited about collaborating via TOPSAN. Our goal is to make TOPSAN easy for them to use, i.e., make sure that they face very little to even no learning curve. Providing transparent access of this nature has three major advantages. Ready usage allows us to benefit immediately from their expertise (in biology), rather than putting them through the grueling task of learning wiki syntax, and the ease of use encourages a larger number of participants so TOPSAN has more content, and becomes more useful. In turn, the more participants, the more each individual participant benefits.

Apart from the website, are there other ways to access TOPSAN?

The free availability of content on TOPSAN, both in terms of license (creative commons share-alike) and online availability allows for efficient reuse of content. We recently integrated TOPSAN into the protein visualization environment on the StarCAVE virtual reality (VR) system at the Calit2 center, UCSD, which allows for the 3D visualization and interaction with protein structural data. Now, in addition to displaying the protein structure we can also display other related information that is available on TOPSAN. This has largely been possible thanks to the Deki API, which allows us to connect to TOPSAN and access content in real-time.

On the StarCAVE, protein structures are projected on the CAVE wall, and one is immersed in a visual environment where these proteins can be extensively manipulated and viewed from many different perspectives. The software was developed by Philip Weber, Andrew Prudhomme (Photo: right), and Jurgen Schulze (Photo: Left front) with some input from me (Photo: Left back).

The StarCAVE consists of five walls with three screens each, and the floor is a screen as well. Our 34 high definition (HD) projectors generate a 360 degrees 3D stereo image. Every projector pair is driven by a computer with an Intel quad core CPU, running under Linux, with dual Nvidia Quadro 5600 graphics cards. We use an additional machine as a head node to control the StarCAVE, making it a total of 18 computers. For head and hand tracking, we use a wireless, optical tracking system with a 3D joystick.

What has the reaction been from your community regarding the TOPSAN project?

TOPSAN has been very well received by the scientific community and our website has been accessed by researchers from all over the world. TOPSAN was recently mentioned in a “News and Views” article on Wikiomics published in a special issue of the journal Nature and in a Nature Reviews Microbiology article on Culture media. TOPSAN has also been described as “a creative and interesting idea with significant potential” by a panel that reviewed the Protein Structure Initiative and as a means for “exploring methods to engage the scientific community in helping to define and refine functional annotation” by the NIGMS director.

What are your future plans for TOPSAN project and how will MindTouch help those initiatives?

While we are still exploring a range of exciting options, it is premature to go into details. However, working closely with the expert engineers at MindTouch, we are actively developing several innovative features that will further facilitate our new scientific communication channel to enable, in turn, the potential for more insight into the actions and roles of novel proteins. Integrating our nascent features within the Deki platform will extend the role of TOPSAN for understanding the dynamic interplay of protein structure and function in health and disease. It will also extend Deki to serve as a model platform for other community-driven processes in communication and discovery.