Title


mrmvivw/mvivw/mvmr -- Multivariable inverse variance weighted regression (MVIVW)

Syntax

mrmvivw/mvivw/mvmr varname_gd varname_gp1 [varname_gp2 ...] [aweight] [if] [in] [, options]

options Description ---------------------------------------------------------------------------------------------- fe fixed effect standard errors (default is multiplicative random effect) gxse(varlist) varlist of genotype-phenotype (SNP-exposure) standard errors level(#) set confidence level; default is level(95) tdist use t-distribution for Wald test and CI limits

Description

mrmvivw/mvivw/mvmr performs multivariable inverse-variance weighted (IVW) regression using summary level data. See Burgess et al. (2015) for more information.

varname_gd variable containing the genotype-disease (SNP-outcome) association estimates.

varname_gp# variable containing the #th genotype-phenotype (SNP-exposure) association estimates.

For the analytic weights you need to specify the inverse of the genotype-disease (SNP-outcome) standard errors squared, i.e. aw=1/(gdse^2).

Options

fe specifies fixed effect standard errors (i.e. variance of residuals constrained to 1 as in fixed effect meta-analysis models). The default is multiplicative random effect standard errors in which case the variance of the residuals is unconstrained and the square root of the estimated residual variance is displayed (Residual standard error). If the residual variance is found to be less than 1 an error message is shown and the model is refitted with it constrained to 1.

gxse(varlist) specifies a varlist of genotype-phenotype (SNP-exposure) standard errors. These should be in the same order as the genotype-phenotype (SNP-exposure) variables in the main varlist. When this option is specified the Q_A statistic for instrument validity is calculated. When this is specified and there are two or more phenotypes conditional F statistics for instrument strength are calculated. See Sanderson et al. (2019) and Sanderson et al. (2021) for more information.

level(#); see [R] estimation options.

tdist specifies using the t-distribution, instead of the normal distribution, for calculating the Wald test and the confidence interval limits.

Examples

Using the data provided by Do et al. (2013).

Setup . use https://raw.github.com/remlapmot/mrrobust/master/dodata, clear

Select observations (p-value with LDL-C < 10^-8) . gen byte sel1 = (ldlcp2 < 1e-8)

Fit MVMR/MVIVW . mrmvivw chdbeta ldlcbeta hdlcbeta tgbeta [aw=1/(chdse^2)] if sel1==1

. mvivw chdbeta ldlcbeta hdlcbeta tgbeta [aw=1/(chdse^2)] if sel1==1

. mvmr chdbeta ldlcbeta hdlcbeta tgbeta [aw=1/(chdse^2)] if sel1==1

Report Q_A statistic and conditional F-statistics . mrmvivw chdbeta ldlcbeta hdlcbeta tgbeta [aw=1/(chdse^2)] if sel1==1, gxse(ldlcse hdlcse tgse)

Stored results

mrmvivw/mvivw/mvmr stores the following in e():

Scalars e(df_r) residual degrees of freedom (with tdist option) e(N) Number of genotypes e(Np) Number of phenotypes e(phi) Scale parameter (root mean squared error) If gxse() specified e(Qa) Q_A statistic e(Qadf) Degrees of freedom of Q_A statistic e(Qap) P-value for Q_A chi-squared test

Macros e(cmd) Command name e(cmdline) Command issued e(setype) Standard error type

Matrices e(b) coefficient vector e(V) variance-covariance matrix of the estimates If gxse() specified e(Fx) vector of conditional F-statistics e(Qx) vector of Q_x statistics

mrmvivw/mvivw/mvmr stores the following in r():

Matrices r(table) Coefficient table with rownames: b, se, z, pvalue, ll, ul, df, crit, eform

References

Burgess S, Dudbridge F, Thompson SG. Multivariable Mendelian randomization: the use of pleiotropic genetic variants to estimate causal effects. American Journal of Epidemiology, 2015, 181, 4, 251-260. DOI

Do et al., 2013. Common variants associated with plasma triglycerides and risk for coronary artery disease. Nature Genetics. 45, 1345-1352. DOI

Sanderson E, Davey Smith G, Windmeijer F, Bowden J. An examination of multivariable Mendelian randomization in the single-sample and two-sample summary data settings. International Journal of Epidemiology, 2019, 48, 3, 713-727. DOI

Sanderson E, Spiller W, Bowden J. Testing and correcting for weak and pleiotropic instruments in two-sample multivariable Mendelian randomization. Statistics in Medicine, 2021, 40, 25, 5434-5452. DOI

Author

INCLUDE help mrrobust-author