Status and Challenges of Reproducibility in Computational Systems and Synthetic Biology

Scientific research is reproducible when the findings can be independently verified. Reproducibility is crucial for the integrity of science. Unfortunately, scientific studies, including computational studies, are often not reproducible. We believe there are two primary causes for the frequent lack of reproducibility of computational systems and synthetic biology studies. First, the information needed to reproduce a result is often not communicated clearly. This issue can be addressed by improving and expanding the existing standards, the support for the standards, and the communication of the standards. Second, the computational environment needed to reproduce a result is often not shared. This issue can be partly addressed with virtual machines. Here, we outline the status of the reproducibility of computational systems and synthetic biology by reviewing the existing standards and software tools. As part of this review, we highlight some of the most common standards and software tools. Additionally, we discuss the shortcomings of the current standards and software tools, highlighting several gaps which continue to make computational systems and synthetic biology studies challenging to reproduce. In particular, we highlight the need for expanded standards for describing the provenance and verification of computational systems biology models.