Personal Genomes Project as a potential EGI community Next Generation Federated HPC infrastructure to drive international genome discovery Peter Walgemoed Carelliance & Dutch Health Hub http://DHH-IPC.nl peterwalgemoed@gmail.com Presented by Ad Emmen Dutch Health Hub & Contrail 1
Community: Personal Genomes Project http://www.personalgenomes.org 2
Information business value Business = Treatment and Research of Breastcancer Congenital Child diseases ….. Information service = Apply knowledge Patient Researcher Apply Data service for to Treatment Treatment and Research and Research Business Value Users: Implementation Experts in treatment/research team Data service = In organisation View+Share (Breastcancer) Business Business Application data case Integration Application1 Application2 local systems Storage <> Data Stewardship IT infrastructure IT infrastructure 3
ARVADOS Open source platform for managing and analyzing biomedical big data Usage catching on in genome community http://arvados.org 4
ARVADOS Open source platform for managing and analyzing biomedical big data Usage catching on in genome community http://arvados.org 4
Challenges 1. Store and organize 100’s of TB’s of large files with multiple meta-data schema 2. Run informatics analyses that do distributed computations on very large datasets 3. Do real-time high-performance queries on compact genome data (e.g. variants) 4. Ensure validity and maintain provenance on all data in the system over time 5. Make it easy to reproduce pipelines exactly as they were done in the past 6. Protect all data with flexible access control rules and strong encryption 7. Share large data sets between data centers and organizations without physically moving data 5
Sharing Real-Time Analysis of Genomic Data “Lightning” Public Cloud etc. Application Framework (APIs and SDKs) Governance Arvados technology Cloud Operating System Distributed Computation “Crunch” Visualization Variant Provenance Private Cloud Analysis Arvados Cloud Diagnostics Cancer “Keep” Data Storage & Management Cloud Arvados Apps
Arvados technology Apps Cancer Variant etc. Diagnostics Visualization Application Framework (APIs and SDKs) Arvados Sharing “Lightning” of Genomic Data Real-Time Analysis Provenance Analysis “Crunch” Computation Distributed Governance Management Data Storage & “Keep” Cloud Operating System EGI Fed Cloud Investigate Integration with EGI federated Cloud Cloud Arvados Cloud Private Cloud Public Cloud
Trusted Digital Repositories E-Discovery Data & Information Services Dutch Breastcancer Data Collection Catalogue Breastcancer Breastcancer Collection Breastcancer Collection Collection UMC Data Hospital Data LRCB Data LRCB Data BC BC B reastcancer B reastcancer B reastcancer Personal Health Record General Practitioner BC Collection Collection Collection BC Data Data HEBON NBCA BVN IKNL CBS Data Data Data Data Data 7
Service Marketplaces Dutch Health Hub start Information Data/app Infrastructure-as-a-Service IT-infra (IaaS) marketplace 8
Some Contrail tools for Dutch Health Hub Market place (RESTful) APIs Gateway Portal Market Place user interface Market Place services Services for Data enrichement servicies (DHH Platform ) specific uses Market Place Data Service Interface IaaS services (computing, storage, networking, ...) HealthHub Data-as-a-service layer 9
Some Contrail tools for Dutch Health Hub Market place (RESTful) APIs Gateway Portal Market Place user interface Market Place services Services for Data enrichement servicies (DHH Platform ) specific uses ConPaaS Web ConPaaS BoT,NoSQL Provider Federated identity SLA Federation management XtreemFS XtreemFS Market Place Data Service Interface IaaS services XtreemFS Cloud file system ConPaaS (computing, storage, SQL, Hadoop networking, ...) HealthHub Data-as-a-service layer XtreemFS 9
Next steps 1.Organise interest in community in Europe 2.Implement test environment as part of Dutch Health Hub 3.Investigate integration with EGI Federated Cloud 10
11
End 11
Recommend
More recommend
