FDR_control_adaptive_imix: The Adaptive Procedure for Across-Data-Type FDR Control for...

In IMIX: Gaussian Mixture Model for Multi-Omics Data Integration

The Adaptive Procedure for Across-Data-Type FDR Control for IMIX Output

Description

The adaptive procedure for across-data-type FDR control based on the output from IMIX models, this can be directly performed by IMIX function, however, if the user is interested in other alpha levels, this function would be useful to avoid rerun the IMIX().

Usage

FDR_control_adaptive_imix(
  imix_output,
  model = c("IMIX_ind", "IMIX_cor_twostep", "IMIX_cor_restrict", "IMIX_cor"),
  alpha
)

Arguments

imix_output	The result output from IMIX() function, result controlled at alpha level only for one component each time.
model	The target model among "IMIX_ind","IMIX_cor_twostep","IMIX_cor_restrict", and "IMIX_cor". Default is IMIX_ind.
alpha	Prespecified FDR control level.

Value

The estimated mFDR for the target component and classify the genes in each component after FDR control at alpha level.

significant_genes_with_FDRcontrol	The output of each gene ordered by the components based on FDR control and within each component ordered by the local FDR, "localFDR" is 1-posterior probability of each gene in the component based on the maximum posterior probability, "class_withoutFDRcontrol" is the classified component based on maximum posterior probability, "class_FDRcontrol" is the classified component based on the across-data-type FDR control at alpha level
estimatedFDR	The estimated marginal FDR value for each component starting from component 2 (component 1 is the global null)
alpha	Prespecified nominal level for the across-data-type FDR control

References

Ziqiao Wang and Peng Wei. 2020. “IMIX: a multivariate mixture model approach to association analysis through multi-omics data integration.” Bioinformatics. <doi:10.1093/bioinformatics/btaa1001>.

Examples

# First generate the data
library(MASS)
N <- 1000
truelabel <- sample(1:8,prob = rep(0.125, 8),size = N,replace = TRUE)
mu1 <- c(0, 5);mu2 <- c(0, 5);mu3 <- c(0, 5)
mu1_mv <- c(mu1[1], mu2[1], mu3[1]);mu2_mv <- c(mu1[2], mu2[1], mu3[1]);
mu3_mv <- c(mu1[1], mu2[2], mu3[1]);mu4_mv <- c(mu1[1], mu2[1], mu3[2]);
mu5_mv <- c(mu1[2], mu2[2], mu3[1]);mu6_mv <- c(mu1[2], mu2[1], mu3[2])
mu7_mv <- c(mu1[1], mu2[2], mu3[2]);mu8_mv <- c(mu1[2], mu2[2], mu3[2])
cov_sim <- list()
for (i in 1:8) {
  cov_sim[[i]] <- diag(3)
  }
data_z <- array(0, c(N, 3))
data_z[which(truelabel == 1),] <- mvrnorm(n = length(which(truelabel == 1)),
mu = mu1_mv,Sigma = cov_sim[[1]],tol = 1e-6,empirical = FALSE)
data_z[which(truelabel == 2),] <- mvrnorm(n = length(which(truelabel == 2)),
mu = mu2_mv,Sigma = cov_sim[[2]],tol = 1e-6,empirical = FALSE)
data_z[which(truelabel == 3),] <- mvrnorm(n = length(which(truelabel == 3)),
mu = mu3_mv,Sigma = cov_sim[[3]],tol = 1e-6,empirical = FALSE)
data_z[which(truelabel == 4),] <- mvrnorm(n = length(which(truelabel == 4)),
mu = mu4_mv,Sigma = cov_sim[[4]],tol = 1e-6,empirical = FALSE)
data_z[which(truelabel == 5),] <- mvrnorm(n = length(which(truelabel == 5)),
mu = mu5_mv,Sigma = cov_sim[[5]],tol = 1e-6,empirical = FALSE)
data_z[which(truelabel == 6),] <- mvrnorm(n = length(which(truelabel == 6)),
mu = mu6_mv,Sigma = cov_sim[[6]],tol = 1e-6,empirical = FALSE)
data_z[which(truelabel == 7),] <- mvrnorm(n = length(which(truelabel == 7)),
mu = mu7_mv,Sigma = cov_sim[[7]],tol = 1e-6,empirical = FALSE)
data_z[which(truelabel == 8),] <- mvrnorm(n = length(which(truelabel == 8)),
mu = mu8_mv,Sigma = cov_sim[[8]],tol = 1e-6,empirical = FALSE)
rownames(data_z) <- paste0("gene", 1:N)
colnames(data_z) <- c("z.methy", "z.ge", "z.cnv")
dim(data_z)

# Fit the model
test2 <- IMIX(data_input = data_z,data_type = "z",alpha = 0.05,verbose = TRUE)

# Adaptive FDR control at alpha 0.2 for IMIX_cor model
fdr_control2 <- FDR_control_adaptive_imix(imix_output = test2, model = "IMIX_cor", 
alpha = 0.2) 

IMIX documentation built on July 14, 2022, 1:05 a.m.