Skip to content

DEPENDENCIES.md

Latest commit

 

History

History
102 lines (79 loc) · 3.83 KB

DEPENDENCIES.md

File metadata and controls

102 lines (79 loc) · 3.83 KB

Software Dependencies — ACTT Validation Study

All analyses are contained in a single Quarto notebook (ACTT_validation_study.qmd) and were conducted in R v4.4.1.

Core Software

Software Version Purpose Source
R 4.4.1 Statistical computing https://www.r-project.org
Quarto ≥ 1.4 Notebook rendering (HTML output) https://quarto.org
ActiLife 6.13.4 GT3X+ data retrieval and export ActiGraph Corporation, Pensacola FL
Act Studio — ActTrust® data download and export Condor Instruments, São Paulo, Brazil

R Packages

Install all packages in R with:

install.packages(c(
  "tidyverse",
  "caret",
  "MuMIn",
  "MetBrewer",
  "broom",
  "cowplot",
  "pROC",
  "reshape2",
  "gt",
  "lmtest",
  "car",
  "msm",
  "effectsize"
))
Package Version Purpose Citation
tidyverse 2.0.0 Data manipulation and ggplot2 visualisation (includes dplyr, tidyr, ggplot2, readr, purrr) Wickham et al. (2019) J Open Source Softw 4(43):1686
caret 6.0.94 Confusion matrix and per-class classification metrics Kuhn (2008) J Stat Softw 28(5):1–26
MuMIn 1.48.4 Multi-model inference and R² computation Barton (2024) CRAN
MetBrewer 0.2.0 Colour palette — "Archambault" used throughout all figures Mills (2022) CRAN
broom 1.0.6 Tidy model output (tidy() for Tukey HSD results) Robinson, Hayes & Couch (2024) CRAN
cowplot 1.1.3 Multi-panel figure composition (plot_grid()) Wilke (2024) CRAN
pROC 1.18.5 One-vs-rest ROC curves and AUC per intensity class Robin et al. (2011) BMC Bioinformatics 12:77
reshape2 ≥ 1.4 melt() for correlation matrix to long format (Figure 2a) Wickham (2007) J Stat Softw 21(12)
gt ≥ 0.9 Cut-point and AUC table rendering in HTML output Iannone et al. (2024) CRAN
lmtest 0.9.40 Breusch-Pagan test for homoscedasticity Zeileis & Hothorn (2002) R News 2(3):7–10
car 3.1.2 Variance inflation factors (VIF) for multicollinearity Fox & Weisberg (2019) Sage
msm 1.8.2 Delta method for 95% CI propagation on nonlinear cut-points Jackson (2011) J Stat Softw 38(8):1–29
effectsize ≥ 0.8 Eta-squared with 95% CI for ANOVA effect sizes Ben-Shachar et al. (2020) J Open Source Softw 5(56):2815

Colour Palette

The notebook uses the "Archambault" palette from MetBrewer with 4 selected indices:

pal <- MetBrewer::met.brewer("Archambault", 11)[c(1, 4, 6, 8)]

These four colours map to: GT3X+ (hip), ACTT (hip), ACTT (wrist), GT3X+ (wrist) respectively throughout all figures. A 5-colour variant is used for Bland-Altman plots:

pal_ba <- MetBrewer::met.brewer("Archambault", 11)[c(1, 3, 4, 6, 8)]

Rendering

To reproduce the full HTML report:

quarto render ACTT_validation_study.qmd

Or open in RStudio (≥ 2022.07) and click Render. The output is a self-contained HTML file (ACTT_validation_study.html) with all figures and tables embedded. Chunk outputs are set to inline.

Notes

  • ActiLife and Act Studio are proprietary tools used only for device initialisation and raw data export. All processing and analysis from that point is performed in open-source R.
  • reshape2::melt() is the only function used from reshape2; it could be replaced with tidyr::pivot_longer() if preferred.
  • The gt package is used for cut-point, equation, and AUC table display in the HTML output and is not required to reproduce statistical results.
  • lmtest, car, msm, and effectsize are used in the supplementary regression diagnostics and reviewer-requested analyses sections.
  • Verify your exact installed versions with packageVersion("package_name") and update the table above accordingly before archiving.