getPosteriorResponseRate: Compute the posterior response rate from MCMC samples

In RGLab/MIMOSA: Mixture Models for Single-Cell Assays

Description

Uses the sampled response indicator when the model class is MCMCResult (or MIMOSAResult encapsulating an MCMCResult) to compute the posterior response rate distribution. This is summarized to its median, 2.5th and 97.5th credible interval.

Usage

getPosteriorResponseRate(x, ...)

Arguments

x	A MIMOSAResultList or MIMOSAResult. All models should be of type MCMCResult, or fitted using method="mcmc" in MIMOSA.
...	unquoted grouping variable name in the pData() table of all the models that specifies the one or more grouping variable by which to compute response rates.

Details

The posterior response rate is the correct way to compare response rates across studies and treatment groups, as it is unbiased compared to the response rate computed from hard thresholds of posterior response probabilities. The credible intervals have the correct behaviour as sample size increases.

Future versions will allow passing of quantiles for summarization, and maybe the full distribution, depending on needs.

Value

a tibble with the grouping variable, and the 2.5th, 25th, median, 75th, and 97.5th percentiles of the posterior response rate.

RGLab/MIMOSA documentation built on Nov. 13, 2020, 5:04 a.m.