Skip to contents

Power for Negative Binomial Mixed Model

Source: R/meff.R
powerNBMM.Rd

Evaluate effective sample size and power for negative binomial mixed model given parameter values

Usage

powerNBMM(N, m, mu, sigSq.a, theta, beta = NA, sd_x = NA, alpha = 0.05, fit)

Arguments

N

number of subjects

m

number of measurements pwer subject

mu

mean read count

sigSq.a

variance of random effect

theta

negative binomial

beta

effect size

sd_x

standard deviation of target variable

alpha

target false positive rate

fit

model fit with fastlmm() or fastglmm()

See also

meff()

Examples

data(PsychAD)

# regression formula
form <- PTPRG ~ (1|SubID) + offset(log(libSize))

# fit NBMM on PTPRG expression
fit <- fastglmm.nb(form, PsychAD)

# Power analysis
powerNBMM(fit = fit, beta = .1, sd_x= 0.5)
#>           N        f   lambda power lambda.asymp.m lambda.asymp.mu
#> SubID 60481 1.234716 186.6921     1       164.1587        185.7482
#>       power.asymp.m power.asymp.mu   N.1        m       mu beta   sigSq.a
#> SubID             1              1 60481 202.9564 1.203237  0.1 0.7856561
#>        theta sd_x alpha   kappa.m  kappa.mu n_measurements   m.mean        rho
#> SubID 0.2445  0.5  0.05 0.8793019 0.9949439            298 202.9564 0.03465127
#>          m.eff    m.max  fraction
#> SubID 25.37575 28.85897 0.8793019