PhenoMeNal: Virtual e-Infrastructure supporting Clinical Research and Metabolism Studies

PhenoMeNal: Virtual e-Infrastructure supporting Clinical Research and Metabolism Studies

The Centre's Dr Philippe Rocca-Serra and Dr David Johnson recently took part in a very fruitful meeting on the Wellcome Trust Genome Campus, in Hinxton, Cambridge, where participants of the EU Horizon 2020 PhenoMeNal project converged for two days of hacking.

PhenoMeNal aims to provide large scale computing for Medical Metabolomics and is led by Professor Christoph Steinbeck of the European Bioinformatics Institute.

Building on the 'Investigation Study Assay' (ISA) model, the main focus of Philippe and David's contribution was to embed the ISA api toolkit to provide reliable experimental metadata tracking, thereby ensuring proper data stewardship, compliance with ethical requirements and capability to deliver long term, effective data preservation to make data reusable and interoperable. In short, it was about establishing a 'FAIR' infrastructure.

Over the course of the hacking event, the Phenome and Metabolomes Analysis (PhenoMeNal) developers deployed, tested and validated a range of workflows to support processes essential to metabolomic signal processing at scales required for modern clinical applications. The core mission of PhenoMeNal is to provide 'spawn and run' cloud compute environments for scientists and clinicians, taking advantage of Google and Amazon web services but also open source cloud solutions such as Open Stack in order to deliver on-demand, scalable resources, which shift shapes to match the compute demands of scientists.

Another major mission of the project is to provide an array of microservices focused on data analysis and arrange them in workflows, building on the Galaxy framework. This framework, well known in the field of genomics and sequence analysis, has now been harnessed to support the Metabolomic Phenotyping use case. The Galaxy workflows mobilize containerized versions (Docker and Biocontainers) of tools such as XCMS for mass spectrometry processing, BatManR for NMR sprectal data processing or Iso2Flux to deal with tracer based metabolomics, thus encompassing the full gamut of techniques used in the field.

In total, 25 apps in phenomenal-h2020/app-library, 32 containers on github.com/phnmnl, 50+ tests on phenomenal-h2020/Jenkins and 5+ test installations, including public.phenomenal-h2020, have been evaluated in preparation of an official release, planned for early Q2 2017.

Follow our team and projects on Twitter:

@Phil_at_OeRC
@NuDataScientist
@PhnmlH2020