Building biomedical web communities using a semantically aware content management system

Sudeshna Das Lisa Girard Tom Green Louis Weitzman Alister Lewis-Bowen Tim Clark Web-based biomedical communities are becoming an increasingly popular vehicle for sharing information amongst researchers and are fast gaining an online presence. However, information organization and exchange in such communities is usually unstructured, rendering interoperability between communities difficult. Furthermore, specialized software to create such communities at low costtargeted at the specific common information requirements of biomedical researchershas been largely lacking. At the same time, a growing number of biological knowledge bases and biomedical resources are being structured for the Semantic Web. Several groups are creating reference ontologies for the biomedical domain, actively publishing controlled vocabularies and making data available in Resource Description Framework (RDF) language.We have developed the Science Collaboration Framework (SCF) as a reusable platform for advanced structured online collaboration in biomedical research that leverages these ontologies and RDF resources. SCF supports structured 'Web 2.0' style community discourse amongst researchers, makes heterogeneous data resources available to the collaborating scientist, captures the semantics of the relationship among the resources and structures discourse around the resources. The first instance of the SCF framework is being used to create an open-access online community for stem cell researchStemBook (http://www.stembook.org). We believe that such a framework is required to achieve optimal productivity and leveraging of resources in - INTRODUCTION Online scientific communitiesgroups of scientists or collaboratories connected through the Internet have become an important means by which to exchange data and information. The most common form of an online community is an intraorganization web site. In this format, a department or a lab, for example, shares data and knowledge in a web-based forum. Barriers to developing a successful community beyond organization or consortia boundaries have been discussed in Bos et al. [1]. The obstacles discussed by these authors include issues, such as scientists preference for working independently and intellectual property competition between institutions. Despite these barriers, the practice of scientists discussing nascent work on the web is an emerging trend, sometimes labeled Science 2.0 [2]. Successful scientific communities in which interdisciplinary researchers network and engage in scientific discussions for a common driving cause have been developed and fill a critical resource gap. One notable example of such a web-based scientific community is Alzforum (www.alzforum. org)a thriving community of over 4600 researchers networking to find a cure for Alzheimers [3, 4]. In Alzforum, researchers can discuss papers and news spontaneously and participate in live discussions. Researchers are also invited to provide perspectives on key research news and comment on papers of the week. Compendia of genes, antibodies, animal models and protocols are also available on the site. Currently, the site contains more than 60 000 literature citations, 1900 research news articles, 6000 comments, 20 000 antibodies, 250 research models, 500 genes from published association studies of late-onset AD, all known mutations causing familial Alzheimer disease, all drugs in Phase 2 and 3 clinical trials and a wealth of community resources, such as databases for grants, conferences and jobs [4]. Another emerging community based on the Alzforum model is the Schizophrenia Forum (www.schizophreniaforum. org)a community of researchers exchanging ideas to develop better understanding of schizophrenia and improve treatment options. Communities such as Alzforum and Schizophrenia Forum require both social and technological infrastructure for nurturing their growth [3, 4]. However, both these sites were evolved over time for their specific communities and until recently there has been no common reusable toolkit to create a new site similar in structure. Data and information in these sites and other similar ones are organized and structured in different ways and there was heretofore only limited opportunity to share and exchange information amongst these sites. Moreover, the use of the Semantic Web [5, 6] to exchange information among these scientific communities in a machine-readable format remains a challenge [7]. At the same time, a large number of biological resources are now becoming available as W3C Resource Description Framework (RDF) triples (http://www.w3.org/TR/rdf-primer/). Gene Ontology (GO), CHEBI and SNOMED [810] are examples of the most widely used ontologies in the biomedical domain. The ambitious BioMoby project that publishes more than 1400 data sources and analysis tools using a semantic framework has also released its first version [11]. The W3C Health Care and Life Sciences Interest Group [12] and other efforts such as, Open Biomedical Ontologies [13] are actively defining common controlled vocabularies and making data available as RDF. One of the goals of Science Collaboration Framework (SCF) is to annotate the discourse, publications and news published within scientific communities with terms and identifiers from these and other semantically characterized biological information resources, and to make the knowledge and linked data available on the Semantic Web. There are other efforts to develop collaborative annotation and knowledge management systems using Semantic wikis [14, 15]. Wikis are being increasingly adopted by the biomedical community for collective annotation. Gene Wiki for collective annotation of gene function [16] is a recently published wiki example. Some of these resources are also available as RDFWikiProteins [17] and BOWiki (http://bowiki.net/). However, wiki is a technology useful for focused annotation efforts and does not easily support community-networking tools such as blogs and forums. Wikis readily enable multiple editing of content and checking the differences between versions. Wiki is a useful technology and is the primary choice when the purpose is to generate a consensus view that is flexible enough to accommodate input from various people. WikiProteins is a great example of that purpose. The entry for human amyloid- A4 protein precursor (APP), http://www.wikiproteins.org/index.php/ Concept:13341741 lists the various functional roles of APP and the types of Alzheimers disease caused by APP defects. However, the provenance of the claims is lost, and the multi-viewpoints, disagreement or divergence regarding the role of APP in Alzheimers are not captured in the entry as it is in Alzforum (http://www.alzforum.org/res/for/jour nal/transcript.asp?LiveID120). The generalist view of APP presented in WikiProteins is not useful to a scientist specializing in Alzheimers research. In summary, wiki is most useful for leveraging the Long Tail and (...truncated)