Simulation (simulate)
June 6, 2026 ยท View on GitHub
Use graphld simulate to generate GWAS summary statistics from the GraphLD simulation model.
Basic Usage
uv run graphld simulate \
output.sumstats \
--num-samples 100000 \
--heritability 0.5 \
--component-variance 1,10 \
--component-weight 0.01,0.001
Unlike the other graphld subcommands, simulate takes an output summary-statistics path as its positional argument rather than an input sumstats file.
Core Options
| Option | Default | Description |
|---|---|---|
-n, --num-samples | Required | Number of samples to simulate |
-H, --heritability | 0.2 | Total trait heritability |
--component-variance | 1.0 | Comma-separated mixture variances |
--component-weight | 1.0 | Comma-separated mixture weights |
--alpha-param | -0.5 | Frequency-dependence parameter |
--random-seed | None | Random seed |
Annotation-Dependent Simulation
The CLI currently exposes annotation loading options, but custom annotation
link functions are only available through the Python API. Use
graphld.run_simulate when you need annotation-dependent effect-size scaling.
| Option | Description |
|---|---|
-a, --annot-dir | Annotation directory |
--annotation-columns | Accepted by the parser, but custom annotation-dependent scaling currently requires the Python API |
--annotation-dependent-polygenicity | Reserved for future support; currently raises NotImplementedError |
The shared options from the CLI overview also apply.