Connecting digital cancer model repositories with markup

Abstract

The cancer research community requires a standardized way of describing mathematical and computational models to enable interoperation between systems, repositories, and between the models themselves. In this paper we describe a new markup language, TumorML, for describing computational models that fall within the domain of cancer. TumorML is an XML-based markup language that wraps existing cancer model implementations with metadata for model curation, parametric interface description, implementation description, and compound model linking. In this paper we first introduce the rationale for a new markup language for computational cancer model description based on our experiences and requirements from the European Commission's 'Transatlantic Tumor Model Repositories' project. The aim of the project is to develop a European-based digital cancer model repository to link and interoperate with a similar established repository based in the United States. TumorML was developed to enable this interoperation between repositories. We introduce the language and describe the main features of the specification and go on to describe a real application of TumorML where a molecular pathway model has been packaged using the new markup language.