multigrain: Optimise Graphical Approaches to Multiple Testing Procedures
Find the optimal graphical multiple testing procedure for
confirmatory clinical trials that test several hypotheses across
endpoints, doses, or subpopulations. Choose the graph maximising an explicit
gain-function: a user-defined measure of trial success that scores each
possible set of rejected hypotheses. Given simulated p-values, search over
hypothesis weights and transition weights to maximise expected trial
success. Conventional power criteria are recovered and the family-wise error
rate is strongly controlled due to the search being confined to graphical
procedures. Described in Spiers, Grayling, Wheeler, and Mander (2026)
"Gain-function optimisation of graphical multiple testing procedures for
confirmatory clinical trials" <doi:10.48550/arXiv.2609.19994>.
| Version: |
1.0.0 |
| Depends: |
R (≥ 4.1.0) |
| Imports: |
cli, dplyr, GA, ggplot2, ggraph, glue, igraph, mvtnorm, nloptr, purrr, Rcpp, RcppParallel (≥ 5.1.10), rlang (≥
1.2.0), stats, tibble, tidyr, utils, vctrs, withr |
| LinkingTo: |
Rcpp, RcppParallel |
| Suggests: |
gMCPLite, spelling, svglite, testthat (≥ 3.0.0), vdiffr |
| Published: |
2026-10-18 |
| DOI: |
10.32614/CRAN.package.multigrain (may not be active yet) |
| Author: |
Alex Spiers  [aut,
cre],
Dragoș Moldovan-Grünfeld [aut],
GlaxoSmithKline Research & Development Limited [cph, fnd] |
| Maintainer: |
Alex Spiers <alex.d.spiers at gsk.com> |
| BugReports: |
https://github.com/GSK-Biostatistics/multigrain/issues |
| License: |
LGPL (≥ 3) |
| URL: |
https://github.com/GSK-Biostatistics/multigrain,
https://gsk-biostatistics.github.io/multigrain/ |
| NeedsCompilation: |
yes |
| SystemRequirements: |
GNU make |
| Language: |
en-GB |
| Materials: |
README, NEWS |
| CRAN checks: |
multigrain results |
Documentation:
Downloads:
Linking:
Please use the canonical form
https://CRAN.R-project.org/package=multigrain
to link to this page.