Progress Report: Ontology-based Annotation of Biomedical Time–series Data

Raimond Winslow and Stephen Granite
The Institute for Computational Medicine
Johns Hopkins University


In a collaborating project with the NCBO, we are developing an ontology for describing the most common form of cardiovascular data – the electrocardiogram (ECG). The ontology includes the concepts needed to describe data collection protocols (number of leads, lead placement, instrumentation), the ECG waveform and its features in health and disease (e.g., QRS, T waves etc; abnormalities of the ECG such as elevated ST segment, inverted T wave, etc), data computed from the ECG (e.g., QT interval variability, QRS width, etc), algorithms used to compute these data, and interpretations (Minnesota codes). It has been deposited at the NCBO (Electrocardiography Ontology, ID 1146). We will review this ontology.
We are developing this ontology so that clinical researchers can manage, analyze, and share semantically labeled ECG data on the CardioVascular Research Grid (CVRG). We will give a demonstration of several components of this data management and analysis workflow. These components have been developed using the Google Web Toolkit, with data visualization supported by the Google Visualization API. The demonstration will include: a) a review of the CVRG Dashboard; b) selection and upload of ECG time-series data; c) data analysis; and d) visualization and annotation of the ECG waveform. The latter is done by invoking Bioportal REST services from the web-interface to retrieve concepts and definitions from the ontology. This represents substantial progress in our effort to develop the “electronic ECG chart recorder”.