bLPM: Fit bLPM

In mingjingsi/LPM: LPM (Latent Probit Model)

Description

Fit bLPM.

Usage

bLPM(data, X = NULL, alpha = NULL, pi1_ = NULL, maxiter = 1e4, tol = 1e-8, coreNum = 1, verbose = FALSE)

Arguments

data	A list of dataframes which contains SNP IDs and p-values for multiple GWASs, the length is the number of GWASs.
X	Design matrix of functional annotations without intercept, where row and column correspond to SNP and annotation, respectively. Default is NULL.
alpha	Initial values of parameter alpha in the Beta distribution which is between 0 and 1 for GWASs, the length is the number of GWASs. Default is 0.1.
pi1_	Initial value of parameter pi1 which is the proportion that eta equals 1 for GWASs, the length is the number of GWASs. Default is 0.1.
maxiter	Maximum number of EM iteration. Default is 1000.
tol	Threshold to stop the EM iteration. Default is 1e-08.
coreNum	Number of cores used. Default is 1.
verbose	Whether or not to output the estimates of parameters in stage 1 and stage2. Defalut is FALSE.

Details

bLPM fits the bLPM. It requires to provide SNP IDs and p-values for multiple GWASs to data, while users can also provide annotation data to X. The number of SNPs for GWASs are not required to be the same.

Value

List of parameter estimation, the value of lower bound of Log-likelihood.

alpha	Estimation of parameter alpha for each GWAS in each pair.
alpha_stage1	Estimation of parameter alpha in the Stage 1 for each GWAS in each pair.
alpha_stage2	Estimation of parameter alpha in the Stage 2 for each GWAS in each pair.
pi1_stage1	Estimation of parameter pi1 in Stage 1 for each GWAS in each pair, which measures the proportion that eta equals 1.
beta0	Estimation of parameter beta for each GWAS in each pair when annotation is not provided.
beta0_stage2	Estimation of parameter beta in Stage 2 for each GWAS in each pair when annotation is not provided.
beta	Estimation of parameter beta for each GWAS in each pair.
beta_stage2	Estimation of parameter beta in Stage 2 for each GWAS in each pair.
rho	Estimation of parameter rho for each pair of GWASs.
R	Reform rho to the correlation matrix R.
L_stage1_List	The value of Log-likelihood in Stage 1.
L_stage2_List	The value of Log-likelihood in Stage 2.
L_stage3_List	The value of Log-likelihood in Stage 3.

Author(s)

Jingsi Ming

See Also

LPM, test_rho

Examples

data(ExampleData)

bLPMfit_first3 <- bLPM(ExampleData$data[1:3], X = ExampleData$X, coreNum = 2)

mingjingsi/LPM documentation built on April 2, 2020, 9:32 a.m.