getBPMCnullmatrix: Generate Monte-Carlo null distributions for a list of...

In nghiavtr/BPSC: Beta-Poisson Model for Single-Cell RNA-seq Data Analyses

Description

Generate Monte-Carlo null distributions for a list of beta-Poisson models

Usage

getBPMCnullmatrix(bp.model.list, fout = NULL, sim.num = 1000,
  useParallel = FALSE, cpu.num = 16, ran.num = 1e+05, E.esp = 0,
  tbreak.num = 10, useDebug = FALSE)

Arguments

bp.model.list	List of beta-Poisson models that are results from estimateBPMatrix function
fout	A *.RData file name to export results
sim.num	A number of simulation of each model
useParallel	An option for using parallel (=TRUE)
cpu.num	The number of cpus if using parallel
ran.num	The number of data points generated from the beta-Poisson model to approximate the theoretical model
E.esp	An small value added to expected value when computing X2, E.esp=0.0 by default
tbreak.num	Number of breaks for binning
useDebug	A parameter setting of getBPMCnull function that is just used for debug and checking, so useDebug=FALSE by default

Value

A list of Monte-Carlo null distributions from the input models (MCdis.list) and setting values of parameters sim.num, ran.num and E.esp

Examples

set.seed(2015)
#create random data matrix from a beta-poisson model
N=10
alp=sample(100,N,replace=TRUE)*0.1;
bet=sample(100,N,replace=TRUE)*0.1;
lam1=sample(100,N,replace=TRUE)*10;
lam2=sample(100,N,replace=TRUE)*0.01;
n=100
bp.mat=NULL
for (i in 1:N)
  bp.mat=rbind(bp.mat,rBP(n,alp=alp[i],bet=bet[i],lam1=lam1[i],lam2=lam2[i]))
#Estimate parameters from the data set
mat.res=estimateBPMatrix(bp.mat,para.num=4,fout=NULL,estIntPar=FALSE,useParallel=FALSE)
MCnullmatrix.res=getBPMCnullmatrix(bp.model.list=mat.res$bp.model.list,fout=NULL,
                                   sim.num=100,useParallel=FALSE)
#Get Monte-Carlo p-values
MC.pval=getMCpval(bp.model.list=mat.res$bp.model.list,
                  MCdis.list=MCnullmatrix.res$MCdis.list)
MC.pval

nghiavtr/BPSC documentation built on May 23, 2019, 4:42 p.m.