Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

GraphTalk Copenhagen - Killing Data Silos in the Life Sciences with Neo4j

Dave Iberson-Hurst, S-cubed
GrapTalk Copenhagen

  • Login to see the comments

  • Be the first to like this

GraphTalk Copenhagen - Killing Data Silos in the Life Sciences with Neo4j

  1. 1. 23/09/2019 1 Killing Data Silos in the Life Sciences with Neo4j Dave Iberson-Hurst Partner, S-cubed 16th September 2019 Copenhagen 2 | ©2019 S-cubed S-cubed • A3 Suite platform • MDR • SWB • Linked Data Services • CDISC Training and Support • Regulatory Development Strategy • Clinical Trial Documentation • Marketing Authorisation Applications & Licence Maintenance • EU SME Status, EU OMPD Holder • QA & GXP Services • Statistical Consultancy • SAS Programming • Data Management • CDISC services • Statistical Analysis and Reporting • Quality Assurance • Biostatistics • Clinical Data Management • Pharmacovigilance •Medical Monitoring • Risk Based Monitoring • Operational Reporting • Qlik Extensions Data Analytics (Qlik) Biometrics Clinical Standards Management Regulatory Affairs
  2. 2. 23/09/2019 2 3 | ©2019 S-cubed Silos 4 | ©2019 S-cubed Silos … but connected
  3. 3. 23/09/2019 3 5 | ©2019 S-cubed Why Clinical Research suffers from silos: o Length of clinical programmes o Number of clinical studies in a programme o Diversity of systems o Organizational Silos o A world of Excel and PDFs o Outsourcing o Complexity o We build in one silo, send it onto the next o We end up with copies of the data for each different purpose o Standards Clinical Research and Silos Protocol Programme Management Data Capture System(s) Standards Regulatory & Safety System(s) Data Warehouse Organisational Boundary 6 | ©2019 S-cubed o Multiple Standards o Different types of standards • Data Exchange (XML) • Content • Terminology o Developed since 2000 o Developed in Silos o Traditionally delivered as PDF o Moving to electronic • MS Excel • XML • RDF o Submission for regulatory approval of products requires use of the standards Data Standards Play a big role in our work
  4. 4. 23/09/2019 4 7 | ©2019 S-cubed Our “Study” World Collect Organize Analyse ResultsPlan 8 | ©2019 S-cubed Our Silos o This world is just too big to understand, too big to fit into our heads o So, as humans, we did what we always do, we cut it up, we decomposed the problem o We developed our standards in silos o As a consequence our data ends up in silos
  5. 5. 23/09/2019 5 9 | ©2019 S-cubed Our Silos o And where we cut we lost the relationship o And so later we use programmers to replace those relationships, stitch the data back together o And each programmer / company potentially does it differently o This causes issues, e.g. data becomes difficult to pool 10 | ©2019 S-cubed Neo4j Property Graph holding multiple study definitions. Links back to triple store definitions Apache Jena/Fuseki Triple Store holding core definitions (e.g. standards) PostgreSQL Relational DB holding users, audit trail etc. Architecture
  6. 6. 23/09/2019 6 11 | ©2019 S-cubed Architecture Baseline Definitions Study Definitions imports 12 | ©2019 S-cubed Benefits o Neo4j holds our study definitions o Use a graph model that replaces the “lost” relationships o Based on solid foundation of industry definitions sourced from the RDF world o Allows for • Scale • Use of third party tools • Take advantage of the large user base • Cross industry use and ideas o The graph allows us to iterate development without significant impact on previous work
  7. 7. 23/09/2019 7 13 | ©2019 S-cubed Benefits o The graph provides for traceability that is required for regulatory submission o Provides for a single source of data o Our required standards become outputs from graph queries reducing the “copies” of data 14 | ©2019 S-cubed Neo4j Use
  8. 8. 23/09/2019 8 15 | ©2019 S-cubed Study Workbench 16 | ©2019 S-cubed Study Workbench
  9. 9. 23/09/2019 9 17 | ©2019 S-cubed Electronic Health Records 18 | ©2019 S-cubed Mining for Definitions
  10. 10. 23/09/2019 10 19 | ©2019 S-cubed And Some Cypher … o We use a lot of rectangular structures, but we can recreate these with Cypher queries 20 | ©2019 S-cubed So … + =
  11. 11. 23/09/2019 11 Contact Details Dave Iberson-Hurst dih@s-cubed.dk

×