MEtaData Format for Open Reef Data (MEDFORD)

Reproducibility of research is critical for science. Computational biology research presents a significant challenge, given the need to track critical details, such as software version or genome draft iteration. Metadata research infrastructures, while greatly improved, often assume a level of programming skills in their user community, or rely on expert curators to ensure that key information is not lost. This paper introduces MEDFORD, a new human-readable, easily-editable and templatable metadata language for scientists to collocate all the details relevant to their experiments. We provide an overview of the underlying design principles, language, and current and planned support infrastructure for parsing and translating MEDFORD into other metadata formats. MEDFORD 0.9 has been specifically designed for the coral research community, with initial metadata generated from RNA-Seq analyses of coral transcriptomes and coral photo collections. Notably, the format is generally applicable and useful for many types of scientific metadata generated by non-computer science experts.