get_prior_ratios: Evaluate prior ratios of MLE estimates

Description Usage Arguments Details Value Note Examples

View source: R/annotation_analysis.R

Description

This function evaluates prior ratios of MLE estimates under user-supplied conditional and marginal prior distributions. If this ratio is above one (or equivalently if the log is above 0), this implies that the conditional prior improves upon the marginal prior for that particular parameter.

Usage

1
2
3
4
5
6
7
get_prior_ratios(
  mpra_data,
  marg_prior,
  cond_prior,
  n_cores = 1,
  verbose = TRUE
)

Arguments

mpra_data

a data frame of MPRA data

marg_prior

a marginal prior

cond_prior

a conditional prior

n_cores

number of cores to use in parallel

verbose

logical indicating whether to print messages

Details

the inputs marg_prior and cond_prior can be taken directly as the outputs of fit_marg_prior and fit_cond_prior.

This output is returned as the difference of log densities.

Value

a data frame of MLE mean and dispersion parameters by variant_id, sample_id, and barcode giving prior ratio for each

Note

The priors for DNA counts are always the same (since we have no prior knowledge of DNA counts)

The current maximum likelihood estimates are sub-par and will be improved. Interpret with caution.

Examples

1

andrewGhazi/malacoda documentation built on Aug. 2, 2020, 12:54 a.m.