CryoEM: From Biomedical impact to Cloud deployment Laura del Caño Jesús Cuenca CNB – CSIC, Madrid
How can one “see” a virus ?
250 nm Confocal optical microscope
0.1 nm (1 Å) X-ray crystallography
From small amounts 0.5 nm of diluted samples, (5 Å) it is possible to solve... ...the structure of large and flexible macromolecular complexes... … without 3D crystals.
“Structural and molecular basis for Ebola virus neutralization by protective human antibodies” Misasi et al. Science 351(6279), 1343-1346. (2016)
MICROGRAPHS MOVIES BIM correction CTF estimation CTF Single Particle Particle Picking Analysis ( SPA ) COORDINATES Preprocessing Postprocessing 2D Classification REFINED 3D Refinement Validation VOLUME 2D CLASSES Estimate resolution 3D CLASSES Initial Model VOLUME 3D classification
Cluster Edition Cloud Edition Desktop Edition Web Tools
Desktop Edition
Web Tools
box
Scipion cluster edition Cluster Edition Universität Basel (Switzerland) NCPS (Shanghai) EPFL (Switzerland) Utah University (USA) IMM (France) Columbia University (USA) Politecnico de Torino (Italy) CIC-Biogune (Spain)
Towards Scipion on the Cloud
SaaS (Software) User data Application Central installation Web access: easy & on demand Billing PaaS (Platform) Libraries Low-level Standard software framework Brick middleware Brick Operating system Built-in scalability and failover Predefined deployment cycle Billing Virtual servers IaaS (Infrastructure) farm Distributed resources: lots of servers and storage across datacenters VM VM VM VM VM (across the world) Elastic computing: dynamic scaling to handle peaks Virtualization layer Resource pooling: real infrastructure is share by all the IaaS users SAN Real servers farm Billing: variable cost (as function of resource use) CLOUD SCENARIOS
Objective “ Lower barriers for scientists to access modern e-Science solutions from micro to macro scales ”. Grid & cloud based infrastructures. Task 2 Cryo-EM in the cloud: bringing clouds to the data. A Competence Center to Serve Translational Research from Mo lecule to Brain.
ENMR.eu VO Deployment for CryoEM processing
Sara HPC Cloud Deployment for CryoEM processing
Sara HPC Cloud Deployment for Instruct training
Objective “ Bring the world of complex data analysis in Structural Biology to a simple Web browser-based Virtual Research Environment .” Task 2 Integration of existing Cryo-EM web services, Scipion Web Tools, on the VRE. World-wide E-infrastructure for structural biology
ENMR.eu VO Scipion Web Tools Deployment
Install and test Scipion on AWS EC2 platform. Create Scipion AMI (not public yet). Test StarCluster (Elastic Cloud images)
Our experience The cloud paradigm is quite different from legacy HPC, but we were able to deploy successfully Remote visualization Cloud architecture for legacy HPC (1.0) Images for cloud (beta)
Challenges Elastic cloud Image publishing to independent repositories Image contextualization Big data transfers Fault-Tolerant High Performance filesystems Security
Our vision: Scipion Ubiquity “ 1-click instances ” in research and commercial clouds: simple provisioning for Scipion showcase & training SaaS : Scipion Web Tools Ready for every client profile / infrastructure: traditional HPC facilities (“owners”) research clouds (paper per use) commercial clouds (pay per use)
Further info “Structural and molecular basis for Ebola virus neutralization by protective human antibodies” Misasi et al. Science 351(6279), 1343-1346. (2016) “Structures of protective antibodies reveal sites of vulnerability on Ebola virus” . Murin et al. PNAS 111(48), 17182–17187. (2014) “Camouflage and Misdirection: The Full-On Assault of Ebola Virus Disease” . Misasi et al. Cell 159(3), 477-486. (2014) “Electron counting and beam-induced motion correction enable near-atomic- resolution single-particle cryo-EM” . Li et al. Nature Methods 10, 584-590. (2013)
Further info Scipion - http://scipion.cnb.csic.es/ INSTRUCT - http://www.structuralbiology.eu/ MoBrain project - https://mobrain.egi.eu Westlife project - http://about.west-life.eu StarCluster - http://star.mit.edu/cluster/ Infrastructure Manager - http://www.grycap.upv.es/im/ Elastic Cloud Computing (EC3): http://servproject.i3m.upv.es/ec3
Acknowledgements Miguel Caballer (UPV) Enol Fernandez (EGI.eu) Boris Parak (CESNET) Nuno Ferreira (SURFsara)
Laura del Caño Jesús Cuenca-Alba ldelcano@cnb.csic.es jcuenca@cnb.csic.es i2pc.cnb.csic.es Follow us on Twitter: @InstructI2PC
Recommend
More recommend
