Big Data and Data Analysis for Personalized Medicine Dr. Paul Terry Ambassador
Agenda Information and Data The Technology The Promise Personalized Medicine 2
CEO/CTO of PHEMI Board of Life Sciences BC Board of Providence Health Care Board of Molecular You Adjunct Professor Big Data SFU 3
1990 San Francisco 49 ers Super Bowl XXIV Edmonton Oilers Stanley Cup Champions West Germany won the World Cup George W Bush Time Person of the Year 1. Cheers (NBC) 2. 60 Minutes (CBS) 3. Roseanne (ABC) 4. A Different World (NBC) 5. The Cosby Show (NBC) 6. Murphy Brown (CBS) 7. Empty Nest (NBC) 8. America's Funniest Home Videos (ABC) 9. The Golden Girls (NBC) 10. Designing Women (CBS) World Wide Web/Internet protocol (HTTP) and WWW language (HTML) created by Tim Berners-Lee. 4
5
BD2K Big Data to Knowledge NIH Initiative (http://bd2k.nih.gov) $32M grants 2014, $656M over next 7 Years 6
311M UKP (524M USD) UK 100k Genomics Project (delivery 2017) 7
Information & Data 8
Information & Data 9
Information & Data 10
Information & Data 11
Information & Data 12
The Technology 13
Big Data Big Data is a Shoe Box of everything 14
Big Data Big Data is a small supercomputer 15
Big Data With a massive (scalable) ingest rate capability June 2013 Published 100M per second 16
Big Data... at a commodity price 17
Big Data... at a commodity price 18
Big Data... for thought 19
Google (+) for Medicine Ambassador Search Maps News Custom Applications Self-Serve Analytics Dashboards Big Data Repository Parse & Index Engines Internet Privacy by Design Analytics-Ready Digital Assets Big Data Repository Parse & Index Engines PHEMI Central Device Image Data Sharing Agreements Silo d Assets Genomics Labs EMR Meds Governance Privacy Security 20
Information Storage Code Fragment xray Database row,cell or entire Gene Seq Webpage EMR Form Consent Letter Documents Virtual machine 21
(Late) Data Modelling Data Feeds Big Data Gather Store Calculate Discovery /Report 22
Digital Library of Assets 23
Your Research Library = Big Data 24
A Digital Asset Digital Object= Raw+ Metadata Timestamp of when it was imported Location/Source imported from Retention policy Classification Data type of asset Version control information Data agreements for the asset Backup information Etc. 25
Digital Library With a small supercomputer acting as Librarian 26
The Promise 27
The Promise of Big Data Schemaless Any Data Type Economics at Scale Data Lake Fast On-Boarding Single Point of Access 50% Lower TCO 28
Ferrari search 29
Data à Information à Knowledge Visualization à Storytelling 30
OLTP
32 OLAP Star Snowflake
33
Data Science 34
Information & Data 35
Taking Big Data Beyond the Data Lake Cataloging Privacy, Security, & Governance Continuously add structure to data Inventory of Digital Assets Keyword, Graph, Geospatial Zero Trust Data Cleanse, Structure, Ontology Data Processing Functions 36
The Promise of Big Data for Personalized Medicine 37
Blue Button 38
Use Cases Autism Wellness Self-Serve Data 39
Autism: Diagnose & Treat Diseases Earlier BioInformatics Tools Analytics Research Portal Family Portal Proprietary & 3 rd party libraries Microarray WES & WGS Physical, birth & medical history Proprietary Database Excel FASTQ, VCF Excel -> EMR SQL Forms Images Metabolome Expression Array Microbiome JSON JPG, DICOM CSV, PDF CSV CSV Link datasets Privacy, security, governance, data sharing agreements, consent management. Replace existing NAS, Spreadsheet, database
Wellness Clinical Chemistry PHEMI Central Big Data Warehouse VCFs & Results Proteome Ambassador Regular molecular screening Early detection Prevent, delay, mitigate Rich research Microbiome Metabolome Lab Reports 41
Self Serve Portal Information on Demand What s the relationship between gene variants & meds for patients with AMI and LVEF<40? Cardiac Outcomes AMI Patients 31# PHEMI Central Big Data Warehouse Ambassador 42 Self Serve Portal
PHEMI Dr. Paul Terry (pterry@phemi.com) Ambassador