The paper “A workflow reproducibility scale for automatic validation of biological interpretation results” is published in GigaScience. The paper is open access and can be found at the following URL: https://doi.org/10.1093/gigascience/giad031

In the field of bioinformatics, the reproducibility of data analysis workflows is a significant challenge. Recent technological advancements have made it possible to reproduce workflow execution, but the evaluation methods for reproduced results are inadequate. This study proposes a “reproducibility scale” based on biological feature values to assess the reproducibility of results. A prototype system was also implemented to automatically evaluate the reproducibility using this scale. Experiments conducted using real research projects and common use cases in bioinformatics demonstrate the effectiveness of the approach. This approach allows for a finer evaluation of reproducibility, considering both result identity and biological interpretation. It enables more specific discussions on reproducibility in bioinformatics.

The software developed under the project is available as an OSS on GitHub. https://github.com/sapporo-wes/tonkaz