Clinical Data Integration For METABRIC

1
Clinical Data Integration For METABRIC
Irene Papatheodorou1, Charles Crichton2, Lorna
Morris1, Peter Maccallum1, Jim Davies2, James
Brenton1 and Carlos Caldas1 1University of
Cambridge, Cancer Research UK Cambridge Research
Institute 2Oxford University Computing Laboratory

• METABRIC
• Molecular Taxonomy of Breast Cancer International
  Consortium
• Analysis of genomic profiling and tissue data
• Clinical data from five different hospitals in
  the UK and Canada
• Clinical data sources use different
  definitions/values/classification schemes for
  similar concepts
• Need for clinical data integration for
  meta-analysis of experiments

• CancerGrid Tools
• ISO11179 Metadata Registry for data element
  registration
• Tools for data Standardization, Querying,
  Inference and Validation (SQIV)
• Produces RDF from XML data and XML Schema with
  SAWSDL attached CDEs (Standardization)
• Queries RDF using SPARQL or converts RDF back to
  XML using CDE marked-up XML Schema (Querying)
• Transforms RDF data into RDF with different CDE
  definitions using pre-defined inference rules
  (Inference)

• METABRIC Database
• Database schema that accommodates the differences
  between data sources
• Each field described by a Common Data Element
  (CDE)
• Database schema translates into five different
  XML schemas

Standardization, Inference Query
XML Data
Inference Rules (JENA)
RDF Data and new CDE IDs
XML Schema new CDE IDs
SQIV Standardization
SQIV Inference
XML Schema CDE IDs
RDF Data and CDE IDs
SQIV Query
SPARQL Query
SQIV Query
XML Query Result
New XML Data