tContrast: tContrast

In BenjaminChittick/BenCScore: Methods For Analysis of High-Throughput Screens

Description

Wraps the tSample function and compares its markov chain to a provided null markov chain. Returns estimated mean, standard deviation, and probability of belonging to the null posterior.

Usage

tContrast(data, mu.prior, var.prior, nu.prior, a, b, null.post, ROPE,
  alternative, runjags.method = "rjparallel")

Arguments

data	activity data for a compound
mu.prior	location parameter for prior distribution of mu, typically the average of all compound activity.
var.prior	spread parameter for prior distribution of mu, typically the square standard error of the mean for all compounds.
a	alpha parameter for prior distribution of compound variance. See EstimateAB.
b	beta parameter for prior distribution of compound variance. See EstimateAB.
null.post	a MCMC sampling of a null posterior with column name mu.
alternative	a character string specifying the relationship to the null posterior, must be one of "two.sided", "greater", or "less".
runjags.method	method by which jags is run, default is rjparallel
rope	Region of Practical Equivalence. This value specifies how close the sampled posterior needs to be to the null posterior to consider them equivalent, default value is 0.

Examples

compound <- rnorm(2, 100, 10)
null.reference <- rnorm(320, 0, 10)
pooled <- c(compound, null.reference)
mu.prior <- median(pooled)
var.prior <- var(pooled)
nu.prior <- 1
a <- 30
b <- 3100
ROPE <- 15
null.post <- tSample(null.reference, mu.prior, var.prior, nu.prior, a, b)
posterior <- tContrast(compound, mu.prior, var.prior, nu.prior, a, b, 
                       null.post, ROPE, alternative='greater')
posterior

BenjaminChittick/BenCScore documentation built on May 5, 2019, 2:41 p.m.