eTRIKS Labs Launch
October 19, 2015
Keith Elliston of the tranSMART Foundation talks Datathons
December 2, 2015

Translational research knowledge management ‘in action’: eTRIKS 2nd Community Meeting

eTRIKS’ second community meeting took place in Amsterdam at the World Trade Centre on the 22nd of October.

The theme of the meeting was “eTRIKS in Action” with a presentation by a key eTRIKS supported project, ABIRISK, and the launch of the eTRIKS Labs initiative.

 

Progress on maximizing the impact of translational research datasets

Scott Wagers CEO of BioSci Consulting.

eTRIKS strives to maximize the impact of  translational research datasets by providing  platforms, services, guidance and trainings that enable efficient integration, staging, exploration, preservation and reuse of translational research data. In our first year we supported 5 projects. Now after 3 years we have supported 27. Our goal is to support at least 40 but the end of the funding period in 2 years.

The initial focus for eTRIKS development team was to contribute to the improvement of the tranSMART platform. The tranSMART platform consists of a tranSMART core which allows for the integration of various types of data, and the tranSMART app which facilitates analytical workflows data exploration.

eTRIKS has contributed substantially to the open source development the tranSMART platform:

Platform Supported data types Analytical workflows
First fully open stack version of tranSMART with Clinical / Phenotypic data ANOVA
PostgreSQL RDBMS Gene expression data Marker Selection
R based analytics Longitudinal Clinical / Phenotypic data Survival analysis
combination of 10 tranSMART code branches miRNA expression Clustering
LDAP integration Proteomics RNAseq analysis
Galaxy plugin Genomics (vcf) aCGH analysis
XNAT plugin RNAseq data Regression
R client library aCGH data Forest plot
Genome Browser
GWAS module


 

eTRIKS is not synomynous with tranSMART

eTRIKS does more to support projects than contributing to the development of tranSMART. The eTRIKS engagement team evaluates projects to understand a project’s needs and helps with:

  • data modeling
  • curation
  • training – over 300 researchers trained
  • system set-up and support
  • access management
  • hosting

eTRIKS has also produced a number of resources

  • guidelines for the secondary use of medical research data
  • standards starter pack for projects
  • guidelines for data curation
  • a public server with over 60 curated datasets

In order to enable scaling and sustainability of the services eTRIKS provides, eTRIKS is in the process of establishing an network of service providers – the eTRIKS Network.

 

Support Case – IMI ABIRISK

Pierre Donnes – CEO at SciCross AB

We have received very good service from eTRIKS. We have not met a problem we couldn`t resolve.” Pierre Donnes – ABIRISK Project

Pierre Donnes provided a rich overview of the IMI and EFPIA supported project ABIRISK, highlighting how eTRIKS has supported their work. ABIRISK is a project that focuses on the mechanisms and consequences of immunization against biopharmaceutical products. With access to retrospective and prospective data of large cohorts across disease domains, ABIRISK aims to improve understanding of anti-drug auto-immunisation and lead the initiative to produce guidelines for drug development and clinical care of patients. To meet these aims eTRIKS has supported ABIRISK’s translational needs. eTRIKS hosts ABIRISK data and has led major efforts to standardise and curate their datasets; applying guidelines set out in the eTRIKS Standards Starter Pack. This has enabled ABIRISK to gain as much value out of their data as possible. Integrated and explorable data are valuable data.

Pierre points out that WP 1 lead by CCN2P3 in Lyon set up and installed a system to manage access to the datasets in the eTRIKS/tranSMART platform. This was one of the aspects of the support that Pierre Donnes highlighted as being crucial. When asked what they would have used in place of the eTRIKS/tranSMART platform and his response was “its likely we`d have resorted to just sending excel sheets back and forth”.

 

Facing up to the legal questions that hinder all major project to project collaborations

The collaboration between eTRIKS and ABIRISK has not been without challenges. The ability to move and access ABRISK data has required the development of a complex material transfer agreement. eTRIKS has driven the development and process of review and ratification of the MTA by each partner of both projects; a legal process that has taken two years. This is a universal problem, which usually drives groups back into an independent research approach. Both eTRIKS and ABIRISK have resisted this.

eTRIKS  partners have been working on a data processing agreement specifically designed to smooth the way for collaboration between major European research projects and reduce the time it takes to enable cross project work on datasets. This initiative has been recognized by the IMI as key to the productive outcome of major European scientific collaborations, and has encouraged eTRIKS to formalise its efforts towards a generally applicable template for data transfer agreements.

 

eTRIKS Labs – A new way to develop ideas in translational informatics

Anthony Rowe – Director, Translational Informatics and External Innovation at Johnson and Johnson

The eTRIKS community also saw the launch of the eTRIKS Labs initiative. The purpose of eTRIKS labs is to cultivate a co-creative environment between developers in the community and users. Giving unprecedented access to ongoing developments in eTRIKS the community can evaluate the suite of applications eTRIKS is working on. The community is free to comment on the developments and critically they are encouraged to contribute technically.

“We want to encourage the landscape to align developmental efforts, to do that we have provided the community access to our technology” – Anthony Rowe

eTRIKS Harmonisation Services (eHS)

Ibrahim Emam of Imperial College London

The eHS will expedite data loading using a new repository for harmonized data and meta-data for translational research projects. We are creating a metadata repository that can be used to standardise study data that is configured using the eHS. The eHS will be connected to tranSMART such that study specific configurations generated within the eHS will enable simple extraction, transformation and loading of datasets into the tranSMART/eTRIKS platform with no, or minimal curator intervention.

Disease Maps

Mansoor Saqi, Alexander Mazein, Irina Roznovat of EISBM

The disease maps are integrated knowledge repositories that provide biological context to data emerging from experimental studies for selected IMI projects. Specifically focussing on respiratory, neurological and rheumatic diseases, this work aims to develop networks that enable the integration of multiple heterogenous biological data sources serving as a database for background knowledge. In addition we are constructing detailed expert-verified metabolic and signaling pathways known to be involved in the disease condition. Ultimately we aim to enable sharing and community curation of these disease knowledge representations, using a variety of frameworks.

SmartR

Sascha Herzinger of the University of Luxembourg

SmartR provides a highly dynamic and interactive way of visualizing and analyzing data within tranSMART. Using recent web technolgies SmartR generates visual analytics within the web browser rather than making use of static images. This enables greater exploration of data without affecting the responisveness of the user interface.

 

The tranSMART Foundation

Keith Elliston – CEO of the tranSMART Foundation

The tranSMART Foundation is a world wide non-profit organisation that drives translational research by means of their very popular knowledge management platform.

“tranSMART progress has been fantastic with the help from eTRIKS and TraIT – Keith Elliston

The tranSMART Foundation can boast a multitude of members including, Oracle, PerkinElmer, Pfizer, BT-Global Services, The Hyve, Philips, Thomson Reuters, IDBS, Harvard University and Imperial College London (to name a few), the most recent member of which being the OpenBEL Consortium.

Open source, open data and open science is what we will mold our future efforts towards, adding to improvements already made in platform reliability, scalability and features; all readily available to our clients.

eTRIKS and tranSMART have worked very closely not only on platform developments, but also on making integrated datasets openly available and usable on the tranSMART Platform. We have now 200 explorable datasets crossing a broad range of disease domains, all available to our clients.

 

Current developments of Imperial College London

Florian Guitton Engineering Lead at Imperial College London

In addition to the eTRIKS Labs, Florian Guitton of the Imperial College of London provided a fantastic summary of the tranSMART platform developments and adaptions they have been working on recently.

  • eTRIKS Analytics Environment – software stack that provides a framework for large scale analytics.
  • XNAT – Open Source imaging informatics software platform
  • High dimensional data retrieval with NoSQL – An efficient means to reduce querying time in tranSMART.
  • Clinical data exploration from federated sources – retrieving data from multiple sources.
  • Integration of the IBM Watson content analytics system – Enables the search of genes and disease associations.
  • tranSMART for large scale visualization suites – a system recently presented first hand to Xi Jinping, President of China.

 

In Summary

Jay Bergeron – Director of Translational and Bioinformatics, Pfizer and eTRIKS Co-Coordinator

After three years we have some very exciting new ideas taking shape. We are eager to see these  developments flourish within our community and wider. We have two years left to generate as much impact as we possibly can, enabling translational researchers across Europe, and the world to draw as much value out of their data as they possibly can.