I2B2-tranSMART hackathon brings I2B2 back as clinical data warehouse

When tranSMART was conceived originally by Recombinant at Johnson & Johnson, it was supposed to extend the clinical warehouse I2B2 with biomarker data, such as microarray gene expression data, and allow the both datatypes to be integrated. TranSMART was designed to use the existing I2B2 architecture as its clinical data storage.

After making the translational research data mart open source one of the first steps the community took, next to making the application work with both Oracle and Postgres, was building a lightweight alternative for I2B2. This alternative implementation of the clinical data warehouse allowed those customers not having I2B2 already deployed to install tranSMART with fewer steps involved.

Over time however, as the alternative became the default way to install tranSMART, tranSMART diverged from its origins and became incompatible with some of the features of I2B2. As I2B2 still has a strong base in hospitals and research institutes around the world the community has recently come together to bring back the cooperation between the two in an I2B2-tranSMART hackathon organized at Harvard University.

Thanks to great efforts of the attending developers, including Gustavo Lopes from The Hyve, representing CTMM TraIT, the support for I2B2 modifiers, samples and time points has been brought back to tranSMART.

The modified tranSMART interface showing (on the left) modifier support and (on the right) support for selecting samples from certain date ranges.

Going forward users again have the choice: installing tranSMART stand-alone or on top of an existing I2B2 installation. Together these tools provide a perfect solution from clinical data storage to translational research on clinical and omics data.