NEW YORK (GenomeWeb) – Researchers in the McDonnell Genome Institute at Washington University School of Medicine in St. Louis have released a beta version of the Clinical Interpretations of Variants in Cancer (CiViC), a public communal resource for capturing, sharing, and discussing information on the clinical significance of genomic variants in cancer.
More specifically, CiViC provides a public forum for sharing written summaries about the clinical relevance of mutations found in sequenced tumor samples. These summaries, which are drawn from published scientific literature, include details about associations between molecular alterations, drugs, and disease prognoses, and more. Members of the community can freely read and comment on posted summaries; suggest corrections, updates, or more nuanced descriptions of variants; submit new evidence statements about particular variants; and even roll these summaries into clinical reports that they generate for clinicians.
It provides a structured mechanism for capturing and publically sharing data on clinically relevant oncology variants that might otherwise stay hidden in clinical reports or research files and slides, Obi Griffith, an assistant professor of medicine in WUSTL's division of oncology, an assistant director in the Genome Institute, and one of CiViC's developers, told GenomeWeb this week.
Not only do posted summaries benefit from the expertise and knowledge that others in the community can provide, but making them available could help reduce duplicated efforts within the community where individual labs develop and use their own internal databases of clinically significant mutations, he said. It'll also free researchers from having to comb through the literature themselves each time they have to generate a report, he added, since they could simply use the summaries provided via the website.
Malachi Griffith, Obi's brother and an assistant director of the Genome Institute and assistant professor in WUSTL's department of genetics, presented CiViC in two talks during the Intelligent Systems for Molecular Biology Conference last week in Dublin, Ireland. In his talk, he said that CiViC is one of a series of WUSTL-developed computational resources for addressing bottlenecks associated with exploring and understanding genomic mutations in various contexts. Others are the Database of Curated Mutations, a repository of known disease-causing mutations that provides variant lists linked to source citations; and the Drug Gene Interaction Database, which aggregates data on drug-gene interactions from multiple scientific databases and published literature making it easier for researchers to search for targeted therapies and potentially druggable genes.
He also said the developers have launched a beta version of the platform and are seeking feedback from the community on the current interface and currently available content. CiViC currently contains over 500 clinical evidence statements focused on 55 cancers, 65 genes, and 140 variants — mostly somatic variants. These statements, generated by a team of about six curators, are based on information provided in about 230 abstracts and papers. As statements build up for a given variant, curators combine them into a single executive summary that describes the clinical relevance of the variant in question.
These initial curation efforts are intended to give people a sense of what their contributions should look like but also provide a baseline on which to build out the CiViC system. Beta testers can contribute to the effort in a number of ways. One of these is simply reading available statements, Malachi said, and identifying errors or ways of restating points to improve clarity. In order to actually contribute content, community members do need to create accounts. Once they do so, they can then suggest new evidence statements for different variants as well as submit edits to already available summaries. Suggested additions go into a queue where they are reviewed by CiViC curators who either approve or reject the statements.
It's similar to the Wikipedia model with the caveat that contributions don't go live until they have been vetted by experts, Obi said. It also shares some DNA with Wikipedia in the sense that contributors don't necessarily have to be biomedical scientists although they should be able to read and understand biomedical literature, he added. "We would love [for example] to have patient advocates get involved," he told GenomeWeb. "They might not have read all the literature but they might know of a new clinical trial that is relevant and they'd be qualified to make an evidence statement about that."
So far, the beta run seems to be going well. Somewhere on the order of 50 to 100 users have signed on to participate, according to the Griffiths. They don't have a clearly delineated end to the beta at this point but the goal is run it for at least a few months giving users an opportunity to kick the tires and provide feedback, they said. Meanwhile, the developers are already working on improvements to CiViC including adding content that they've generated as well as bulk imports from collaborators at institutions such as the Genome Sciences Center in Vancouver, the developers of the ClinGen database, and members of the Global Alliance for Genomics and Health, who are interested in adopting CiViC for their curation tasks.
The developers are also mulling ways to handle community contributions as they come in. Curation is both time- and energy-consuming and so it is important to find ways to make this more efficient, particularly as the number of contributors grows. One of the benefits of collaborating with other groups is that curators on those other teams could help evaluate contributions to CiViC once they've become familiar with the system, the Griffiths said. There may also be opportunities for contributing community members who consistently provide good content and comments to move into more editorial roles, they said. This model has certainly worked for other bioinformatics resources such as the BioStars forum. In addition to beefing up the content, CiViC's developers are also working on improvements to the user interface including making it easier to use and incorporating new helper features.
Malachi noted during his talk at ISMB that CiViC is similar to a number of existing commercial solutions as well as systems developed and maintained by researchers at various academic centers and hospitals. It's probably most closely aligned with the My Cancer Genome database, which was developed by researchers at Vanderbilt University's Ingram Cancer Center and provides information on the clinical significance of genetic mutations associated with multiple cancer types and their impacts on treatment; and the Personalized Cancer Therapy resource developed by researchers at MD Anderson, he said. CiViC also competes in some sense with the National Center for Biotechnology Information's ClinVar database, which aggregates information about genomic variants and links to human health.
But CiViC fills in some of the gaps left by these existing resources, according to the Griffiths. For the first two, the main difference is openness of the model, Malachi said. The data in those systems comes from an in-house team of experts and there really aren't clear mechanisms for external researchers to make contributions. In contrast, a big part of CiViC is its user interface and the ease of accessibility by the larger scientific community, he said. Moreover, CiViC offers an application programming interface that facilitates computational access to variant summaries, a feature that those other databases do not have, he added.
ClinVar is a little different, Obi said. Its emphasis is on "collecting variants that are being observed in the wild in the sense that if you sequence a patient and see a specific variant ... and you have some information about its pathogenicity or not, you can upload that variant into ClinVar," he explained. "Certainly it's linked to clinically relevant information but it isn't focused on the curation of the literature ... which is what we are focused on."
Malachi noted as much during his talk at ISMB in which he listed a few things that CiViC is not intended for. These include using it as a forum for providing information on variants of unknown significance which may or may not be clinically relevant or for posting unpublished findings — although published case reports are currently accepted under a special category, he said.