Package 'MVMR.CUE'

Multivariable Mendelian Randomization Accounting for Correlated Horizontal Pleiotropy

Description

Multivariable Mendelian randomization accounting for complex correlated horizontal pleiotropy while elucidating shared genetic etiology.

Author(s)

Qing Cheng [email protected]

---

Example MVMR Summary Statistics

Description

Example summary statistics for multivariable Mendelian randomization, containing association estimates and standard errors for 200 genetic variants with two exposures and one outcome.

Usage

data(gamma1)
data(Gamma2)
data(se1)
data(se2)

Format

gamma1

A 200 x 2 matrix of SNP-exposure association estimates.

Gamma2

A matrix of 200 SNP-outcome association estimates.

se1

A 200 x 2 matrix of standard errors for SNP-exposure associations.

se2

A matrix of 200 standard errors for SNP-outcome associations.

---

Generate Raw Genetic Data

Description

Simulates SNP genotype data with a specified minor allele frequency range and correlation structure.

Usage

genRaw(L, M, rho, n)

Arguments

L	Number of independent LD blocks.
M	Number of SNPs per LD block.
rho	Correlation parameter controlling linkage disequilibrium within blocks.
n	Sample size.

Value

An $n \times (L \times M)$ matrix of simulated genotype values (0, 1, or 2).

Examples

## Not run: 
    X <- genRaw(L = 10, M = 5, rho = 0.3, n = 500)
  
## End(Not run)

---

HORSE Mendelian Randomization via MCMC

Description

Implements the HORSE (Heteroscedastic Outlier-Robust causal effect estimation with Shrinkage for plEiotropy) model for univariate (mr_horse) and multivariable (mvmr_horse) Mendelian randomization using JAGS.

Usage

mr_horse(D, no_ini = 3, variable.names = "theta", n.iter = 10000,
         n.burnin = 10000)
mvmr_horse(D, no_ini = 3, variable.names = "theta", n.iter = 10000,
           n.burnin = 10000)

Arguments

D	A data frame containing columns betaY, betaX, betaYse, betaXse (and betaX1, betaX1se, betaX2, betaX2se, etc. for mvmr_horse).
no_ini	Number of MCMC chains (default 3).
variable.names	Character vector of parameters to monitor (default "theta").
n.iter	Number of MCMC iterations per chain after burn-in (default 10000).
n.burnin	Number of burn-in iterations (default 10000).

Value

A list with elements MR_Estimate (a data frame of point estimates and diagnostics) and MR_Coda (the MCMC samples as a mcmc.list).

Examples

## Not run: 
    D <- data.frame(betaY = Gamma2, betaYse = se2,
                    betaX = gamma1, betaXse = se1)
    res <- mr_horse(D)
  
## End(Not run)

---

Multivariable LASSO Mendelian Randomization

Description

Implements multivariable MR with LASSO-based invalid instrument selection. cv.mvmr_lasso performs cross-validation to select the LASSO penalty parameter; mvmr_lasso uses this to obtain post-LASSO causal estimates.

Usage

mvmr_lasso(bx, by, seby)
cv.mvmr_lasso(bx, by, seby)

Arguments

bx	Matrix of SNP-exposure association estimates (SNPs x exposures).
by	Vector of SNP-outcome association estimates.
seby	Vector of standard errors for SNP-outcome associations.

Value

cv.mvmr_lasso returns a list with elements fit (LASSO coefficients at the selected lambda) and lambda.

mvmr_lasso returns a list with elements thest, a, lambda, th_post, and se_post.

Examples

## Not run: 
    res <- mvmr_lasso(bx = gamma1, by = Gamma2, seby = se2)
  
## End(Not run)

---

Multivariable Weighted Median Mendelian Randomization

Description

Implements multivariable weighted median MR using quantile regression. mvmr_med_boot computes bootstrap standard errors.

Usage

mvmr_median(bx, sebx, by, seby, boot = FALSE, boot_it = 1000)
mvmr_med_boot(bx, sebx, by, seby, N)

Arguments

bx	Matrix of SNP-exposure association estimates (SNPs x exposures).
sebx	Matrix of standard errors for SNP-exposure associations.
by	Vector of SNP-outcome association estimates.
seby	Vector of standard errors for SNP-outcome associations.
boot	Logical; if TRUE, bootstrap standard errors are computed.
boot_it	Number of bootstrap iterations (default 1000).
N	Number of bootstrap iterations in mvmr_med_boot.

Value

mvmr_median returns a list with element coefficients and, if boot = TRUE, element se.

mvmr_med_boot returns a vector of bootstrap standard errors.

Examples

## Not run: 
    res <- mvmr_median(bx = gamma1, sebx = se1, by = Gamma2, seby = se2)
  
## End(Not run)

---

Multivariable Robust Regression Mendelian Randomization

Description

Implements multivariable MR using robust regression (MM-estimator) with inverse-variance weighting.

Usage

mvmr_robust(bx, by, seby, k.max = 500, maxit.scale = 500)

Arguments

bx	Matrix of SNP-exposure association estimates (SNPs x exposures).
by	Vector of SNP-outcome association estimates.
seby	Vector of standard errors for SNP-outcome associations.
k.max	Maximum number of iterations for the M-step (default 500).
maxit.scale	Maximum number of iterations for the scale estimation (default 500).

Value

A list with elements coefficients and se.

Examples

## Not run: 
    res <- mvmr_robust(bx = gamma1, by = Gamma2, seby = se2)
  
## End(Not run)

---

Multivariable Mendelian Randomization Accounting for Correlated Pleiotropy

Description

Core functions implementing the MVMR-CUE algorithm for multivariable Mendelian randomization that accounts for complex correlated horizontal pleiotropy.

Usage

MVMRCUE(gammahM, Gammah, se1M, se2, Re, opts = NULL)
MVMRCUEfun(gammahM, Gammah, se1M, se2)
MVMRCUEoverlapfun(gammahM, Gammah, se1M, se2, Re, Eta)
MVMRCUEIndepSample(gammahM, Gammah, se1M, se2, opts = NULL)

Arguments

gammahM	Matrix of SNP-exposure association estimates.
Gammah	Vector of SNP-outcome association estimates.
se1M	Matrix of standard errors for SNP-exposure associations.
se2	Vector of standard errors for SNP-outcome associations.
Re	Correlation matrix for overlapping samples.
Eta	Eta parameter for overlapping samples.
opts	Optional list of algorithm options.

Value

A list containing the estimated causal effects and associated statistics.

Examples

## Not run: 
    result <- MVMRCUE(gamma1, Gamma2, se1, se2, Re = diag(nrow(gamma1)))
  
## End(Not run)

---

Set of functions in example RcppArmadillo package

Description

These four functions are created when RcppArmadillo.package.skeleton() is invoked to create a skeleton packages.

Usage

rcpparma_hello_world()
rcpparma_outerproduct(x)
rcpparma_innerproduct(x)
rcpparma_bothproducts(x)

Arguments

x	a numeric vector

Details

These are example functions which should be largely self-explanatory. Their main benefit is to demonstrate how to write a function using the Armadillo C++ classes, and to have to such a function accessible from R.

Value

rcpparma_hello_world() does not return a value, but displays a message to the console.

rcpparma_outerproduct() returns a numeric matrix computed as the outer (vector) product of x.

rcpparma_innerproduct() returns a double computer as the inner (vector) product of x.

rcpparma_bothproducts() returns a list with both the outer and inner products.

Author(s)

Dirk Eddelbuettel

References

See the documentation for Armadillo, and RcppArmadillo, for more details.

Examples

x <- sqrt(1:4)
  rcpparma_innerproduct(x)
  rcpparma_outerproduct(x)

---