Reproducible Research

Our approach to reproducible research consists of three main components:

1. Data Sharing (repositories)
2. Audit Trail (version control systems)
3. Reproducible Analysis (literate programming)

The idea is very simple aligns well with the principles of the scientific paradigm. For each experiment or study whether you are a basic scientist or a quantitative researcher following these guidelines will vastly improve reproducibility. Reproducibility in the laboratory is fraught with difficulties and although much of this web resource is dedicated to quantitative scientists, there are key concepts that can be helpful for wet-lab scientists. In fact, the first two components require the participation of the those producing the data. Although, laboratory experiments may be difficult to reproduce at times for a long list of reasons, we believe that there are no excuses when for lack of reproducibility when it comes to that analysis portion of a study.

The main reasons that the vast majority of scientific studies do not employ the above components or comparable steps are several. First, it is not the common practice—researchers tend to follow the habits of their mentors and colleagues. Second, there is no established on one size fits all approach to reproducible research, there is a tendency to hold on to the data even after publication, and lastly the learning curve for implementing these steps can be daunting.

Our aim in this web resource is to promote these three steps as a generalized approach to reproducible research by providing resources to reduce the difficulty of the learning curve

Data Sharing

Resources

• Data Sharing

Audit Trail

Resources

• Audit Trail

Reproducible Analysis

Resources

• Data Sharing
• Audit Trail
• Reproducible Analysis

External Resources