The story of an experiment: A provenance-based semantic approach towards research reproducibility.

Keywords

Authors

Abstract

Endtoend reproducibility of scientific experiments is a key to the foundation of science. Reproducibility of an experiment does not necessarily guarantee the accuracy of its results, but it guarantees that the steps of an experiment can be repeated to a certain level of significance to generate similar results. Data provenance plays a key role in telling the story of an experiment which helps one step towards reproducibility. To convey the message of a story, it is essential to provide sufficient data and its flow along with its semantics. In this paper, we present a provenancebased semantic approach to explain the story of a scientific experiment with the primary goal of reproducibility. The REPRODUCEME ontology extended from PROVO and PPlan is used to represent the whole story of an experiment describing the path it took from its design to result. We visualize and evaluate the provenance lifecycle of a scientific experiment taking into account the use case of life science experiments.