Your first analysis

Run a complete pathway-first example before preparing your own files. The maintained script uses 80 synthetic samples, two bundled signatures, GLM and one requested Response endpoint. No AI service or API key is needed. The data are synthetic.

Locate the example

library(IOBRcausal)
script <- system.file("examples", "pathway_first_minimal.R", package = "IOBRcausal")
stopifnot(nzchar(script), file.exists(script))
rscript <- file.path(R.home("bin"),
  if (.Platform$OS.type == "windows") "Rscript.exe" else "Rscript")
Sys.setenv(IOBRCAUSAL_EXAMPLE_OUTPUT = file.path(path.expand("~"), "iobrcausal-demo"))

You can also download the same English script. It uses the public iobrcausal() entry for the complete workflow.

Preview, then run

# Preview the inputs and planned stages.
system2(rscript, shQuote(script))

# Execute the synthetic example when you are ready.
system2(rscript, c(shQuote(script), "--run"))

The preview does not fit models. The executed script prints the actual delivery directory inside ~/iobrcausal-demo (your home folder). Open that printed path to find the results. If several package libraries exist, use the same R installation and check the engine path rather than assuming the subprocess loaded your latest package.

Open the result

Open README.md inside the delivery directory printed by the script. Follow its links to validation, the endpoint report, full statistics and figures. Read the actual stage statuses; an empty candidate sheet can be a valid outcome.

What to inspect	What it tells you
Requested endpoint	Only Response was requested in this example
Workflow stages	Which steps completed, were empty, skipped or failed
Formal estimate table	Effect scale, uncertainty, method and support
Pathway-gene mapping	Which original members connect each pathway to genes
Full audit tables	Rows hidden from a short display, including failures

To adapt this to your files, continue with data preparation and the pathway-first workflow. To inspect an existing delivery, use saved results.