Power Analysis for Differential Expression Studies
Add confidence intervals for power estimates
Add Bayesian posterior predictive intervals for power estimates
depower: Power Analysis for Differential Expression Studies
Test statistic distribution under the null
Evaluate confidence intervals for power estimates
Evaluate Bayesian posterior predictive intervals for power estimates
GLM for NB ratio of means
GLMM for BNB ratio of means
GLMM for Poisson ratio of means
Likelihood ratio test for BNB ratio of means
Likelihood ratio test for NB ratio of means
MLE for BNB
MLE for NB
Negative log-likelihood for BNB
Negative log-likelihood for NB
Plot power objects
Simulated power
Simulate BNB data
Simulate log-transformed lognormal data
Simulate NB data
Paired and one-sample t-Tests
Welch's t-Test
Wald test for BNB ratio of means
Wald test for NB ratio of means
Provides a convenient framework to simulate, test, power, and visualize data for differential expression studies with lognormal or negative binomial outcomes. Supported designs are two-sample comparisons of independent or dependent outcomes. Power may be summarized in the context of controlling the per-family error rate or family-wise error rate. Negative binomial methods are described in Yu, Fernandez, and Brock (2017) <doi:10.1186/s12859-017-1648-2> and Yu, Fernandez, and Brock (2020) <doi:10.1186/s12859-020-3541-7>.