Contents
associate$mr
Mendelian-randomization entry point for a PolyGeniusStudy
associate$mr() reserves the Mendelian-randomization surface. It validates
data and its arguments and then always aborts: no estimator backend is
implemented. The signature is published so that calling code and documentation
can be written against a stable shape.
Usage
associate$mr(
data,
exposure,
outcome,
instruments = c("prs", "variants"),
method = c("ivw", "egger", "median"),
covariates = NULL,
weights = NULL,
output = c("summary", "models"),
...
)Arguments
| Argument | Description |
|---|---|
data | A PolyGeniusStudy. |
exposure | Unquoted exposure expression, resolved from data. Required. |
outcome | Unquoted outcome expression, resolved from data. Required. |
instruments | One of "prs" (default), "variants". Instrument type the call would use. |
method | One of "ivw" (default), "egger", "median". MR estimator the call would use. |
covariates | Unquoted adjustment expression, or NULL (default). Captured and discarded. |
weights | Unquoted expression or numeric vector, or NULL (default). Captured and discarded. |
output | One of "summary" (default), "models". Captured and discarded. |
... | Reserved for future MR backends. Captured and discarded. |
Value
Never returns. Always raises a not-yet-implemented error after validating the requested interface.
Details
Validation runs in order: data must be a PolyGeniusStudy; exposure and
outcome must both be supplied and non-NULL; instruments, method and
output are matched against their choices. covariates, weights and ...
are accepted and discarded. The abort message echoes the resolved
instruments, method and output so a caller can confirm the interface the
call would have requested.
See Also
associate, associate$regression
Other associate-analyses:
associate-compare,
associate-mediation,
associate-meta,
associate-regression,
associate-single-variant