TranSMART Platform Powers International Asthma Research Initiative

By Bio-IT World Staff 

October 1, 2014 | The U-BIOPRED project began with a big goal: to understand the factors that define severe asthma cases, splitting the disease into meaningful subtypes that may let doctors form better prognoses and treatment strategies. Ultimately, the project hopes to discover biomarkers that could be used to quickly diagnose subtypes of asthma, or even lead to new therapies. To do that, researchers are combining data from longitudinal clinical studies of both adults and children, and animal, in vitro and in silico models of the disease. Under the U-BIOPRED banner, a battery of multi-omics technologies is producing an unprecedented wealth of data on the cellular changes that underlie severe asthma.

U-BIOPRED, or Unbiased BIOmarkers in PREDiction of respiratory disease outcomes, is a project of the Innovative Medicines Initiative (IMI), the largest European public-private partnership in the field of medicine, with a specific goal of laying the groundwork for new pharmaceuticals. The IMI invested €20 million ($25.4 million) in U-BIOPRED, distributed to a consortium that includes 20 academic institutions, ten members of the commercial European Federation of Pharmaceutical Industries and Associations, six patient advocacy groups, and others.

Aligning all of those stakeholders, each contributing different types of data, presented a host of IT challenges. U-BIOPRED required a centralized platform that could store, integrate, and analyze clinical and multi-omics data, while offering secure access to all members and providing a venue to collectively manage hypotheses and publications. The successful adoption of tranSMART v1.1 has allowed the project to progress without bottlenecks around collaboration and data-sharing, producing early research results and earning U-BIOPRED a 2014 Bio-IT World Best Practices Award in Research and Drug Discovery.

The Data Challenge

The tranSMART platform originated at Johnson & Johnson, but is now overseen by an independent tranSMART Foundation that seeks to reduce barriers between research organizations by supporting a single, open source data platform for a multitude of biomedical data types. (See, “tranSMART Launches Demo with BT, Membership Drive.”) U-BIOPRED members had access to an early proof-of-concept version, tranSMART v0.9, and found it was able to handle both clinical and transcriptomic data fluidly. “At the beginning of the project, tranSMART was one of the only platforms capable of hosting data as complex and variable as that planned to be produced in the consortium,” says U-BIOPRED contributor Ioannis Pandis of Imperial College London.

As the consortium prepared to integrate data collected from around 1,000 patients with severe asthma, the release of tranSMART v1.1, which included support for new types of omics data and expanded access control, ensured its adoption as U-BIOPRED’s central platform.

The project’s first task was to analyze clinical and omics data from around 700 adult and 300 pediatric asthma patients. Clinical data for each patient was collected over three visits, each of which saw around 650 variables per patient entered into an electronic Case Report Form. Samples were then profiled with transcriptomic, genomic, proteomic, and lipidomic technologies. U-BIOPRED also introduced a category of measurements its members call “breathomics,” which identify metabolites in samples of exhaled air collected from patients. Today, U-BIOPRED members have collected nearly 90 million individual data points from patients and various preclinical models of asthma. With all this information stored centrally and accessible across institutions, it can now form the basis for new discoveries about the biomarkers that define asthma cases with shared causes and molecular signatures.

So far, the scientific returns on investment have been very encouraging, say the authors of U-BIOPRED’s entry in the Bio-IT World Best Practices competition. Ten papers are currently in the publication pipeline, and data has also been contributed to the UK Asthma Registry via a simple tranSMART download.

A Window for Collaboration

While tranSMART contained the tools U-BIOPRED needed to store, analyze and share its data, a platform for higher-level collaboration was not initially available. Members could see the same data, but lacked a central forum to discuss their findings and next steps.

To facilitate cooperation within U-BIOPRED, the Imperial College of London, a leading member of the consortium, developed the Knowledge Portal. This wiki-based platform provides a hub that contains both raw data and the metadata associated with it, analysis results, tranSMART training materials, and information that project leaders upload outlining goals and measures of progress in their research. The Knowledge Portal also records hypotheses and analysis plans, which members can reference to see where separate studies may intersect and inform one another. An important feature of the Knowledge Portal is tracking the provenance of each data submission, providing a transparent trail as numerous organizations contribute to the U-BIOPRED project.

Combining tranSMART with the Knowledge Portal for collaboration has allowed the U-BIOPRED consortium to expedite data harmonization, access and analysis in the context of a large-scale international research project in biomarker discovery. This is a model that could be imitated by other multi-stakeholder collaborations.

“The complexity of current research projects dictates the use of appropriate data and knowledge management platforms, in order for a project to be successful,” says Pandis. “Interestingly, we are starting to see a swift change in funding body and researcher attitude towards the use of such platforms.”

The success of the tranSMART implementation for U-BIOPRED has inspired IMI to invest €23 million ($29 million) in a cloud knowledge management project, to enable new research initiatives to store and share study data on a common platform. The European Translation Information and Knowledge Management Services (eTRIKS) project currently supports, maintains and extends tranSMART for the U-BIOPRED project, as well as a number of EU and UK-based translational medicine projects established by IMI and regional funding bodies. Tools first created for U-BIOPRED may find expanded use as eTRIKS seeks best-in-class knowledge management platforms to serve as the basis for a growing universe of European research programs.

“The eTRIKS project, which is a major partner of the tranSMART Foundation, is currently looking into ways of abstracting the structure of the Knowledge Portal to develop it as a solution for all eTRIKS supported projects, plus potentially offer it as a tranSMART plugin,” says Pandis.

Meanwhile, as U-BIOPRED focuses increasingly on mining understanding from its massive database, the tranSMART Foundation continues to innovate. This September, the Foundation launched tranSMART v1.2, with new search and retrieval functions, and a toolset to help tranSMART users create their own extensions to the platform.

The U-BIOPRED consortium is already preparing to take advantage of the upgraded platform, as the project scales up its study populations. “The recent features included in tranSMART v1.2 will enable the hosting and analysis of GWAS [genome wide association study] data, which has only recently been produced by the consortium,” says Pandis. This phase of U-BIOPRED research will search for heritable genetic mutations that predispose individuals to develop asthma — and may also shed light on the molecular pathways underlying the disease.