Nsim_stat: Nsim_stat

In jamesliley/subtest: Test whether two putative subgroups of a disease phenotype show evidence of differential genetic basis.

Description

Computes how well the parameters gamma and kappa are fitted at various numbers of simulations, and how these affect the computation of p-values. Returns an object of class 'sim_N_test' which can be plotted.

Usage

Nsim_stat(lr, NN = c(50, 100, 500, 1000, 1500), NB = 500, p_at = c(2, 3,
  5, 10), quant_at = pnorm(-2:2), verbose = TRUE)

Arguments

lr	matrix of simulation results
NN	vector of integer; test performance of simulation fit for n simulations, where n takes the values in NN
NB	number of bootstrap samples to take for each sample size in NN
p_at	compute p values at each value in p_at
quant_at	for each size in NN and each parameter, record these quantiles of the bootstrapped parameter
verbose	print values of NN[i] as they complete.

Value

list of two objects, stat and quant; stat[,ii,i] is, in order, kappa, gamma, and p-values given gamma and kappa of each value in p_at for the i'th bootstrap sample for bootstrap sample size NN[ii]; quant[k,ii,j] is the quant_at[j]'th quantile of boostrapped test statistic k amongst bootstrap samples of size NN[ii]

Author(s)

James Liley

Examples

data("sim_mat.RData") 
stat=Nsim_stat(sim_mat)
summary(stat)
plot(stat)

jamesliley/subtest documentation built on May 18, 2019, 11:21 a.m.