FdrInfo: Estimate FDR thresholds based on permuted outcome label...
In pjnewcombe/Pmisc: A collection of miscellaneous functions.

Description Usage Arguments Value Author(s)

Given observed posterior probabilties, and those calculated under permutations of the outcome labels, this function calculates posterior probability thresholds for different target false discovery rates.

FdrInfo(observed.post.probs = NULL, results = NULL,
  permuted.post.probs = NULL, perm.res.files = NULL, predictors = NULL,
  target.fdrs = c(0.01, 0.05, 0.1), n.cuts.order = 5,
  diagnostic.plot.info = TRUE)

`observed.post.probs`	A vector of posterior probabilities from the observed (i.e. ‘real’) data analysis. Note - MUST be named. Either this must be provided, or the results object below.
`results`	Java MCMC results class object. Either this must be provided, or the vector of posterior probabilities above.
`permuted.post.probs`	A matrix of posterior probabilities from permuted analyses - rows are variables and columns are the different permuted outcome analyses. Either this must be provided, or the list of results files below.
`perm.res.files`	A list of results files for all the permuted outcomes analyses with identifier "_perm1" included in the file name for, e.g., the first permutation. If multiple imputation was carried out the results file corresponding to each imputation is included in the list. Either this must be provided, or the matrix of posterior probabilities from each permuted label analysis (see above).
`predictors`	Which subset of predictors to perform the FDR calculation over. Default NULL means all will be considered.
`target.fdrs`	A vector of FDR levels to calculate posterior probability thresholds for. Default is 0.01, 0.05, 0.1
`n.cuts.order`	Order of magnitude for vector length of possible thresholds to explore
`diagnostic.plot.info`	Whether to generate information for a diagnostic plot (involves repeating calulcations adding each permutation one by one) which can be time consuming. Default it TRUE.

fdr.info A list containing the target FDR rates, info for each including the posterior probability threhsold and the estimated FDR rate at that threshold across the permutations. The list of covariates reaching each threshold is also included

Paul Newcombe

pjnewcombe/Pmisc documentation built on March 26, 2020, 2:09 p.m.