Skip to main content

BioMedical Information Integration (MedII) for Pharmacogenomics

Healthcare & Life Sciences

Technical Overview

BioMedical Information Integration - MedII (formerly Shaman, IMR) helps realizing the Information Based Medicine vision and includes three layers.

The first is an integration layer that includes tools and transformations from propriety formats to standard XML-based representation such as the emerging HL7 Clinical Document Architecture (CDA) for clinical documents, CDISC Operational Data Model (ODM) for clinical trial data, CDISC Study Data Tabulation Model (SDTM), OMG Microarray Gene Expression (MAGE), and Bioinformatic Sequence Markup Language (BSML). It also includes the UDiP technology to de-identify according to HIPAA the data prior to leaving its source premises.

The index layer creates two indices for the data - a text-based index and a relational database index enabling MedII to support structured, semi-structured, and unstructured data and facilitating composite mining of the data using free-text search (google-like), XQuery and SQL.

The EHR and Capsules layer (Electronic Health Record layer), intended for application in the healthcare domain, creates patient-centric objects that are compiled from the documents indexed by the previous layer.

Mining and query tools can reside on top of either the second layer, as in the case of pharmacogenomics, or on top of the third layer. MedII provides one such novel data mining tool - FANO (and ProDoc), which is based on multivariant analysis for studies of positive/negative associations and covariances. It facilitates complex queries such as: 'What protocols were used for tumors that produced similar staining sections and was from same age group as my patient, and with same "Yakamura" polymorphism in her genes?'

MedII is used in IBM Clinical Genomics solution.