The function proceeds as follows:

1. Select features (by default this is done using LASSO feature selection).

2. Subset the mvdat to only retain relevant features and instruments.

3. Perform MVMR on remaining data.

mv_subset(
mvdat,
features = mv_lasso_feature_selection(mvdat),
intercept = FALSE,
instrument_specific = FALSE,
pval_threshold = 5e-08,
plots = FALSE
)

## Arguments

mvdat Output from mv_harmonise_data. Dataframe of features to retain, must have column with name 'exposure' that has list of exposures tor etain from mvdat. The default is mvdat_lasso_feature_selection(mvdat). Should the intercept by estimated (TRUE) or force line through the origin (FALSE, the default). Should the estimate for each exposure be obtained by using all instruments from all exposures (FALSE, default) or by using only the instruments specific to each exposure (TRUE). P-value threshold to include instruments. The default is 5e-8. Create plots? The default is FALSE.

List of results