Webinars | Susheel Varma
As a Scientific Workflows Coordinator for a €15m EC-funded project called VPH-Share, my day-to-day work involves coordinating collaborative development of software, services, workflows and institutional processes to help translate scientific research into patient-specific treatment plans.
I’m intimately involved with the software architecture, development and deployment of all technical and scientific projects, including deploying large-scale sensitivity and uncertainty analysis tools on hpc and cloud compute farms. I also help coordinate scientific and technical delivery of 25+ projects associated a various levels of maturity and healthcare domains.
Previously, for the VPH-NoE project, I was involved in the evaluation, design and deployment of federated data storage and workflow technologies for the VPH.
My doctoral work at the Kroto Research Institute, Sheffield, involved building large agent-based models of cardiac cells and tissues, where I also helped and contributed to the development two open-source projects, Flame & Chaste.
18th May 2017 (14:00 – 14:45 GMT) : “The VPH-DARE@IT Platform for Translating Research to Clinical Decision Support of early and differential diagnosis of dementia”
The VPH-DARE@IT project  is funded under the EU’s 7th Framework Programme and involves 20 partners across eight countries. It aims at improving early and differential diagnosis of dementia, by means of a unique and multi-faceted image-based phenomenological analysis and mechanistic modelling paradigms for personalised evaluation of patients and forecasting. The phenomenological approach relies on analysing large multi-centre cross-sectional and longitudinal datasets looking for statistical patterns and relationships to determine potential biomarkers [2,3] for differential diagnosis of Dementia or for the early identification of cognitive decline. The mechanistic approach models the underlying physiological mechanisms of diseases, proceeding from fundamental biophysical principles  to forecast the pathophysiological evolution of subject in time and identify potential new biomarkers.
The project has converged on building two software platforms (see Figure 1): a Clinical Research Platform designed to foster integration and testing of new datasets, tools and workflows across the consortium and a Patient Care Plat-form aimed at distilling the resulting knowledge and providing a clinical decision support tool targeted at specialists and general practitioners. This talk will outline the integrated bimodal IT approach undertaken by the project consortium to leverage Cloud Computing and HPC services to integrate over 16+ datasets [5,6,7,8,9], 83+ scientific data processing tools [10,11,12,13] and 15+ workflows  (see Figure 2), which has eventually resulted in processing over 6000 anonymised patient records using 15,000 workflow executions in 200,000 CPU hours run solely within the pub-lic cloud infrastructure in two months. The next challenge is expanding the platform to incorporate machine learning services to process 500,000 patient data from the UK Biobank.The talk will summarise the VPH-Dare@IT project’s comprehensive platform to provide clinical “Research As A Service” (RaaS) enabling open innovation and interactions between a community of users in unprecedented ways. Hardware & data providers, scientific software developers, clinical researchers and clinicians have all benefited with the development of a community of practice around an advanced research ecosystem that enables all parties to grow, nurture and sustain research and translate ideas into products, services and vertical platforms for the precision medicine community.
1. The VPH-Date@IT project web page. http://www.vph-dare.eu.
2. Jack CR et al., The Lancet Neurology. 2013;12(2):207-216.
3. Young AL et al., Brain. 2014;137(9):2564-77.
4. Vardakis JC et al., Computer Models in Biomechanics. 2013;305-316.
5. Mueller SG et al., Neuroimaging Clin N Am. 2005;15(4):869-877.
6. Weiner MW et al., Alzheimer’s & Dementia. 2012;8(1):S1-S68.
7. Ellis KA et al., International Psychogeriatrics. 2009;21(04):672-687.
8. Bateman RJ et al., New England Journal of Medicine. 2012;367(9):795-804.
9. Visser P et al., Neuroepidemiology. 2008;30(4):254.
10. FreeSurfer. http://freesurfer.net
11. CONN: https://www.nitrc.org/projects/conn/
12. C. Pierpaoli et al. Tortoise ISMRM; 2010 (1597)
13. Nipype: Neuroimaging in Python Pipelines and Interfaces http://nipy.org/nipype/
14. Coveney P et al. (eds.), Computational Biomedicine, Oxford University Press, 2014