Skip to contents

Test mediator-instrument independence of negative controls (W | Gm | C)

Source: R/nc_diagnostics.R
nc_independence_check_gm.Rd

For each control feature, computes the partial correlation with the mediator-specific genetic instrument Gm after residualising both on observed covariates C, and reports the p-value. Controls significantly associated with Gm after FDR correction may carry eQTL / allele-specific effects that violate the mediator-instrument-independence assumption (A2'). This is the mediator-instrument analogue of nc_independence_check(): just as the exposure instrument G must be independent of the negative controls, so must the mediator instrument Gm.

Usage

nc_independence_check_gm(dat, fdr_level = 0.1, n_cores = 1)

Arguments

dat

Dataset list from run_single_iteration() or generate_toy_data(), containing Gm, W, and synthetic_data. If Gm is absent (i.e. no mediator instrument was generated), the function returns NULL with a message.

fdr_level

Target FDR for BH correction. Default 0.10.

n_cores

Number of parallel workers. Default 1 (sequential). Uses parallel::mclapply on Unix and a PSOCK cluster on Windows.

Value

A data frame with one row per control feature: feature, partial_r, p_value, fdr, significant, verdict. Returns NULL if dat$Gm is not present.

Details

For a mediator cis-eQTL instrument, this screen tests whether the eQTL instrument is associated with the negative-control panel – a violation would indicate shared genomic structure between the instrument SNPs and the control features.

Examples

dat <- run_single_iteration(n_features = 10, phi = 0.8, seed = 1)
nc_independence_check_gm(dat)
#> NC independence (Gm): 10 tasks (sequential)
#>  NC independence (Gm): 10% (1/10) [0s]
#>  NC independence (Gm): 20% (2/10) [0s]
#>  NC independence (Gm): 30% (3/10) [0s]
#>  NC independence (Gm): 40% (4/10) [0s]
#>  NC independence (Gm): 50% (5/10) [0s]
#>  NC independence (Gm): 60% (6/10) [0s]
#>  NC independence (Gm): 70% (7/10) [0s]
#>  NC independence (Gm): 80% (8/10) [0s]
#>  NC independence (Gm): 90% (9/10) [0s]
#>  NC independence (Gm): 100% (10/10) [0s]
#>    feature    partial_r   p_value       fdr significant verdict
#> 1        1 -0.005942095 0.8946642 0.9568965       FALSE   valid
#> 2        2 -0.002425623 0.9568965 0.9568965       FALSE   valid
#> 3        3  0.014979713 0.7385298 0.9568965       FALSE   valid
#> 4        4 -0.003474061 0.9382972 0.9568965       FALSE   valid
#> 5        5 -0.019308759 0.6669895 0.9568965       FALSE   valid
#> 6        6  0.003262888 0.9420410 0.9568965       FALSE   valid
#> 7        7  0.011658359 0.7950306 0.9568965       FALSE   valid
#> 8        8 -0.009218868 0.8372426 0.9568965       FALSE   valid
#> 9        9  0.003499733 0.9378421 0.9568965       FALSE   valid
#> 10      10  0.002933582 0.9478818 0.9568965       FALSE   valid