Reproducible Research

CautionUnder construction

• Reproducibility is defined as obtaining consistent results using the same data and code as the original study (synonymous with computational reproducibility).
• Replicability means obtaining consistent results across studies aimed at answering the same scientific question using new data or other new computational methods.

Reproducibility and Replicability in Research

Use version control (git or svn)

• Do not track model development in git, it is too messy, which messes with the git history.
  • Use rsync if needed
  • Track the Rmd-file for the report
  • This tracks the models. The models are still in the “messy” folder.
    • base_model <- run25.mod
    • covariate_model <- run63.mod
    • final_model <- run67.mod
    • simulation_model <- run68.mod
  • Runrecord
    • runno
    • based on
    • OFV
    • dOFV
    • Condition number (CN)
• Do not track produced PDFs in git

When and how to commit

Commit often and in small, contained chunks.

The seven rules of a great Git commit message¹

1. Separate subject from body with a blank line
2. Limit the subject line to 50 characters
3. Capitalize the subject line
4. Do not end the subject line with a period
5. Use the imperative mood in the subject line. Git itself uses the imperative whenever it creates a commit on your behalf. A properly formed Git commit subject line should always be able to complete the following sentence: If applied, this commit will your subject line here
6. Wrap the body at 72 characters
7. Use the body to explain what and why vs. how

File/folder naming

• Name files/folders using only A-Z, a-z, 0-9, -, _.
  • Start folder names with a number for sorting purposes.
• In general, use kebab-case for naming (easier to read than snake_case).
  • If there are multiple parts to a name (e.g., a description, a date, and an author), use snake_case to separate between parts, and kebab-case within the parts (e.g., descriptive-name_2025-01-08_viktor-rognas.ext)

Folder structure:

project/
  - README.md       # Project description
  - input/
    - data/         # All input data files
      - raw_data/   # Untouched original data files
        - raw_data.csv
      - dat1.csv
      - dat2.csv
  - R/              # R-scripts
      - dat1.R
      - dat2.R
  - NONMEM/
    - model/        # Model files
      - pk/
        - run001.mod
      - pd/
        - run002.mod
  - output/         # Results
    - report/
      - 1a/
        - .tex
        - .pdf
      - 1b/
        - .tex
        - .pdf
      - 1/
        - .tex
        - .pdf
    - presentation/ # Communication
      - slides.pptx

Coding: language specific

R

• Script all plots.
• Quarto-scripted report.
  • R.version
  • rstudioapi::versionInfo()
  • .packages()
  • devtools::session_info(pkgs = "attached")

NONMEM

When using Monte-Carlo estimation methods (e.g., SAEM, IMP, or FOCE MCETA)), always specify the SEED option and RANMETHOD=P. Also, it is recommended to specify the RANMETHOD option accordingy: * For SAEM and IMP: RANMETHOD=3S2P * For MCETA:RANMETHOD=4P($SIMULATION` uses this method by default)

Footnotes

1. https://cbea.ms/git-commit/