Page 1 of 2   next page »

This document describes an application currently being constructed with Windows Presentation Foundation (WPF) and Office SharePoint Server 2007 (OSS) that will enable collaboration between cancer researchers called the C-ME project (Collaborative Molecular Modeling Environment). Office 2007 with Visual Studio Tools for Office v3.0 (VSTO) will be used to generate reports from the information stored in OSS. The C-ME project is being built for cancer researchers. Curing cancer is a complex and lengthy process, and the collaborative tools researchers need don't exist. This document describes the problems of collaboration faced by cancer researchers and the technical approaches being taken to bridge the gap.

Introduction
The Scripps Research Institute (TSRI), one of the country's largest, private, non-profit research organizations, has been at the forefront of basic biomedical science, a vital segment of medical research that seeks to comprehend the most fundamental processes of life. In just three decades the Institute has established a track record of major contributions to the betterment of health and the human condition. The Kuhn-laboratory at TSRI is translating basic science into bedside therapeutics and diagnostics for patients by bringing together people and technology. Research is focused on cancer diagnostics and emerging infectious diseases.

The Challenge
Several discussions revealed distinct challenge areas where technology could enable a higher level of productivity and collaboration. For example, the research and analysis that occurs over weeks, months, and sometimes years gives birth to a detailed molecular model. To the layman, the model looks like a blob of atoms. However, to the researcher, the model might reveal the secrets to curing cancer. Spinning, twisting, and turning the molecules in a virtual graphical environment reveals the smallest of atomic details that spawn hundreds of discussions, dozens of white papers, and further research. Groups across the world can view the molecules and share complex thoughts and ideas. But without the proper tools, collaboration can only occur in pockets, discussions are lost in e-mail threads, and ideas and experiments are re-hashed because they aren't effectively collected. Good collaborative tools tuned to the needs of scientists, research technologists, and doctors simply don't exist.

The Vision
To enable more effective collaboration, information has to be accumulated differently. Mechanisms to generate content already exist, from stream of consciousness thinking to more formal documentation, but there's no good way to organize and later find this molecular metadata. For example, the molecular model reveals to a researcher an unusual interaction between two atoms. A discussion regarding this detail would ensue in the form of an e-mail thread with a small group of colleagues. Synergy starts, ideas are shared, but it quickly ends after some days pass and the knowledge is eventually lost in an inbox. And even if the researcher decided to save this e-mail thread for later discovery, the context of the thread is lost. All discussions and documentation need to stem from the molecular model to provide context, and the molecular model must be used as the map to the metadata.

The Solution
A solution that lets users collaborate win the context of the molecular model would significantly aid researches in the effort to cure cancer. The technologies chosen to build the C-ME project were WPF, Office 2007, and SharePoint v3.0. WPF would allow for a graphically intensive application that would render the molecules with their associated metadata. The graphical environment would let researchers navigate and collaborate in ways that were never before feasible. OSS would be leveraged to store all of the collected information. In effect, the new application would be a highly contextualized visualization of the information stored in OSS. Office 2007 would be used to collect information, and VSTO would automatically enable reporting from the collected data stored in OSS.

In the solution a layer of collaboration object code will be structured so that it can be shared across different types of presentation layers. In phase 1 of the C-ME project, a custom WPF application that can load protein database (PDB) files and images will be created (Figure 1). In addition, the out-of-the-box Web Services provided by OSS v3 will be used for all interactions. The side effect of programmatically storing information in OSS is that OSS searches and navigation can still be used to mine the information.

The current design for the SharePoint folder structure involves dynamic creating an OSS site for each collaborative hotspot. A hotspot on an image is a region of the image, but in a molecule it's one or more selected atoms. The collaborative sites will be created under a site managing the molecule PDB file. In each site, multiple document libraries will be automatically created. InfoPath form libraries are used whenever possible to collect data in a structured manner. For example, the notes metadata type is a simple InfoPath form. Document thumbnails are essentially image libraries with metadata properties.

The collected information, which is the metadata to the molecular model, manifests itself in different forms. Each type of metadata is a direct mapping to how data is currently collected. The strategic difference is that now all the metadata will have been given context automatically by using the solution (Figure 2).

Current Metadata
Data is generated in various ways, including hallway conversations, e-mails, white papers, and lab books. However, there's currently no tool or method that organizes this information. OSS alone only provides a repository of information, but the structure is manually created within it. Without a tremendous amount of vigilance to constantly create context for every piece of data, it's nearly impossible to sort through the volume of information collected during the course of research.

Currently, OSS, e-mail, lab notebooks, and many meetings are the tools that are used. This is also the model that most businesses run on today. The inability to quickly categorize and organize the information that everyone lives with is exacerbated by the amount of information being collected. The proposed solution is a metaphor for how other industries can give context to their data as a by-product of how they do their work, and not an afterthought. For example, a document is associated with an atom in a specific molecule. The act of associating this document to a specific location would automatically populate properties of the document.

New Metadata
Different types of molecular metadata will be collected. These types map to existing forms of collaboration. It's not the goal to create new forms of collaboration, only to tie it together in ways never conceived before.

Attachment Association
The concept of attaching a metadata type isn't exclusive to the atoms of a molecule. Each metadata type can be attached to another. For example, one researcher might attach a note to an atom and then a discussion ensues from the note. Additionally, a group of atoms can be selected and an attachment can be associated. Finally, metadata is also collected at the level of the entire molecule; so facilitating attachment at the molecular level is necessary.

Discussions
Discussions occur in many forms. If you're not physically together and are in different offices, you use the phone, instant messages, blogging, e-mail, attachments, and workspaces. In contrast, if you're together, you could use the former and also face-to-face communications (the water cooler, for instance.). All forms of this kind of collaboration can be characterized as focused stream of consciousness thinking, and typically they are never cataloged in a way that others can use.

The blog metaphor works well for this kind of collaboration. Blog threads can be attached to any point of the molecular model. Multiple discussion threads that all researchers can contribute to are accessible from the model. Anytime new responses are generated, the researcher can opt to get an e-mail but will always be graphically notified throughout the model where new responses have been created.

Notes
Notes are collected in many different ways, say, from a yellow sticky pad or lab book to your computer. A note is something you can write for yourself or for others. It can serve as a personal reminder or a simple communication to another person, but no extensive discussion has to occur beyond it. A note should be quickly writeable and you should be able to easily stick it anywhere to the molecule to give it context.

A researcher should be able to apply a sticky note that's either personal or public to any part of the molecule. All notes will be graphically organized by author, but can be quickly filtered to sift through the volumes of information. A discussion thread could be generated from a sticky note.

Documents
A variety of documents can be generated at any time. These are stored on file shares, a researcher's workstation, and even SharePoint. The folder that the document lives in and/or the name of the document provides minimal context. A document is any unstructured piece of data, typically created with Office.

Just like you can attach a document to an e-mail, you'll be able to attach a document to any part of the molecule. A SharePoint document workspace will be created behind-the-scenes to store the document. When other researchers load the molecular model, the new attachments will be visually apparent.

Document Thumbnails
Researchers simply use too much information to expect all of it to be maintained in a single repository. The ability to reference a document is required. The reference could manifest in a URL to the associated document, and be stored as metadata to an atom, group of atoms, or other metadata types. However, because the volume of information is so vast, a simple reference isn't enough. For example, if the referenced document is a 50-page PDF file, it's simply not feasible to provide adequate context as to what part of the document was of interest when the association was formed. The problem is compounded when dozens of large documents are referenced.

Page 1 of 2   next page »