generate_data_be: Simulate summary statistics according to the binary-effects...

In dan11mcguire/mamba: MAMBA: Meta-analysis Model Based Assessment of Replicability

Description

Simulate summary statistics according to the binary-effects (be) model.

Usage

generate_data_be(M = 50 * 10^3, p = 0.01, tau2 = 0.00025,
  resref = NULL, n = rep(50 * 10^3, 10))

Arguments

M	number of SNPs for which to simulate summary stats.
p	proportion of SNPs with real non-zero effects.
tau2	variance of real associated non-zero SNPs.
resref	an optional vector of residual variances to sample from (with replacement) when generating the standard errors for the summary stats. Default is NULL, in which case the residual variances calculated from a Cigarretes Per Day GWAS (Liu, Jiang 2019) are used.
n	vector of sample sizes from the contributing studies.

Value

A list containing:

• An Mxk matrix of effect size estimates betajk,

• An Mxk matrix of effect size estimate variances sjk2,

• M-length vector inverse-variance weighted meta-analysis z-scores meta.z,

• an M-length binary vector indicating real / non-real effect at each SNP Rj,

• an M-length binary vector indicating true effect-size at each SNP muj.

• an M-length vector indicating the number of studies at each SNP with non-zero effects .

• an Mxk binary matrix indicating which studies at which SNP had non-zero effects.

Examples

generate_data_mamba(M=100, p=0.01, tau2=2.5e-4, n=rep(10^4, 10), lambda=0.975, alpha=15) 

dan11mcguire/mamba documentation built on Nov. 10, 2020, 12:37 a.m.