iProMix: Cell-Type Specific Association Analysis for Bulk Profiling Data

Description Usage Arguments Value Examples

For Y matrix with multiple genes/proteins/variables, calculate the likelihood ratio test statistics (as well as chisq p-value) for iProMix

iProMix.LRT.matrix(
  yMatrix,
  x,
  cov = NULL,
  pi,
  reduce1 = c(2, 1),
  reduce2 = NULL,
  tuningPar = 1e-06,
  cl = FALSE
)

`yMatrix`	The quantitative measure (e.g. protein/expression) of a gene in the matrix form (Row: Genes; Columns: Samples)
`x`	The quantitative measure of anther gene (e.g. ACE2 protein levels) that we would like to know their cell-type specific dependency with Y
`cov`	The covariates for adjustment. Their impact on the mean value of X and Y are adjusted
`pi`	The proportion of cell type 1
`reduce1`	A index of the row and column of the variance-covariance matrix that should be forced to be zero in cell type 1. Default is NULL.
`reduce2`	A index of the row and column of the variance-covariance matrix that should be forced to be zero in cell type 2. Default is NULL.
`tuningPar`	Default is 1e-8. It is used in the embedded graphic lasso procedure for estimating correlation. A larger tuningPar can be selected if one is interested in penalized estimates.
`cl`	True of False. If true, parallel computing is used; need the library(doRNG). Default is FALSE

list with number of elements equal to the number of genes for the yMatrix. Each gene contains a sublist with 10 elements. It contains

`var1:`	The estimated (y,x) variance of cell type 1
`var2:`	The estimated (y,x) variance of cell type 2
`mu1:`	The estimated (y,x) mean function of cell type 1
`mu2:`	The estimated (y,x) mean function of cell type 2
`cor.score1:`	The estimated X-Y correlation in cell type 1
`cor.score2:`	The estimated X-Y correlation in cell type 2
`full.ll:`	The estimated log likelihood function from the full model
`reduced.ll:`	The estimated log likelihood function from the reduced model
`LRT:`	The likelihood ratio test statistics for iProMix
`LRT.pvalue:`	The chisq p-value for iProMix

library(iProMix)
set.seed(111)
y <- matrix(rnorm(100,10,1), ncol=10)
x <- rnorm(10,10,1)
pi <- runif(10)
iProMix_result <- iProMix.LRT.matrix(yMatrix = y, x = x, pi = pi, reduce1=c(2,1), reduce2=NULL)