A simulated summary-statistics example for the IBMR workflow with one
exposure, one primary outcome, and two candidate auxiliary outcomes. The data
are simulated (not real GWAS data): the exposure has a genuine positive effect
on the primary outcome, Auxiliary_1 shares pleiotropic (invalid-instrument)
structure with the primary outcome and is therefore informative for instrument
borrowing, while Auxiliary_2 shares little and is not.
Usage
data(ibmr_example)Format
A named list with the following entries:
- BetaXG
Numeric vector of SNP-exposure associations.
- seBetaXG
Standard errors for
BetaXG.- BetaYG_matrix
Matrix of SNP-outcome associations; columns
Primary,Auxiliary_1, andAuxiliary_2.- seBetaYG_matrix
Standard errors matching
BetaYG_matrix.- primary_trait
"Primary".- candidate_auxiliaries
c("Auxiliary_1", "Auxiliary_2").- recommended_auxiliary
"Auxiliary_1".- dat_ibpresso_aux1
Data frame formatted for
IBPRESSO()with the primary outcome and the recommended auxiliary outcome (Auxiliary_1).- source_note
Description of how the example was generated.