Table 3 Guidelines for reproducible research and for the development of high-quality methods.

From: Ensuring scientific reproducibility in bio-macromolecular modeling via extensive, automated benchmarks

General guidelines for reproducibility

Guidelines for high-quality benchmarks

1. Document artifacts

1. Define scientific questions for the benchmark

2. Share input, output, and exact workflow in detail under an open license in public repositories

2. Define quality metrics that are practically relevant

3. Cite the data, software, and workflows

3. Diversify examples in the benchmark set to cover easy and difficult targets

4. Use persistent links in the publication

4. Separate benchmark set from the developed method

5. Journals should check for reproducibility

5. Pick cutting edge methods to compare your method to

6. Funding agencies should fund reproducibility research

6. Use benchmarked methods that are freely available

Back to article page