new_control: Control parameters for the BayesProt Bayesian model

In biospi/bayesprot: Bayesian Protein-level Quantification for Proteomics

Description

Control parameters for the BayesProt Bayesian model

Usage

new_control(feature.model = "independent", feature.eb.min = 3,
  peptide.model = NULL, peptide.eb.min = 3,
  assay.model = "independent", assay.eb.min = 3,
  error.model = "poisson", missingness.model = "censored",
  missingness.threshold = 0, model.seed = 0, model.nchain = 4,
  model.nwarmup = 256, model.thin = 1, model.nsample = 1024,
  squeeze.var.func = squeeze_var, dist.var.func = dist_invchisq_mcmc,
  dist.mean.func = dist_lst_mcmc,
  nthread = parallel::detectCores()%/%2, hpc = NULL)

Arguments

feature.model	Either single (single residual) or independent (per-feature independent residuals; default)
feature.eb.min	Minimum number of features per peptide to use for computing Empirical Bayes priors
peptide.model	Either NULL (no peptide model; default), single (single random effect) or independent (per-peptide independent random effects)
peptide.eb.min	Minimum number of peptides per protein to use for computing Empirical Bayes priors
assay.model	Either NULL (no assay model), single (single random effect) or independent (per-assay independent random effects; default)
assay.eb.min	Minimum number of assays per protein protein to use for computing Empirical Bayes priors
error.model	Either lognormal or poisson (default)
missingness.model	Either zero (NAs set to 0), feature (NAs set to lowest quant of that feature) or censored (NAs modelled as censored between 0 and lowest quant of that feature; default)
missingness.threshold	All feature quants equal to or below this are treated as missing (default = 0)
model.seed	Random number seed
model.nchain	Number of MCMC chains to run
model.nwarmup	Number of MCMC warmup iterations to run for each chain
model.thin	MCMC thinning factor
model.nsample	Total number of MCMC samples to deliver downstream
nthread	Number of CPU threads to employ
hpc	Either NULL (execute locally), pbs, sge or slurm (submit to HPC cluster) [TODO]

Value

bayesprot_control object to pass to bayesprot

biospi/bayesprot documentation built on Nov. 9, 2019, 2:40 p.m.