DGU: Differential gene usage in immune repertoires
In snaketron/test_repo: Differential gene usage in immune repertoires

Description Usage Arguments Details Value Author(s) See Also Examples

View source: R/Dgu.R

IgGeneUsage detects differential gene usage in immune repertoires that belong to two biological conditions.

1
2
3

DGU(usage.data, mcmc.warmup, mcmc.steps,
    mcmc.chains, mcmc.cores, hdi.level,
    adapt.delta, max.treedepth)

`usage.data`	Data.frame with 4 columns: 'sample_id' = character identifier of each repertoire, 'condition' = character key representing each of the two biological conditions, 'gene_name' = character name of each gene to be tested for differential usage, 'gene_usage_count' = number of rearrangements belonging to a specific sample_id x condition x gene_name. Alternatively, usage.data can be a SummarizedExperiment object. See examplary data 'data(Ig_SE)' for more information.
`mcmc.chains, mcmc.warmup, mcmc.steps, mcmc.cores`	Number of MCMC chains (default = 4), number of cores to use (default = 1), length of MCMC chains (default = 1,500), length of adaptive part of MCMC chains (default = 500).
`hdi.level`	Highest density interval (HDI) (default = 0.95).
`adapt.delta`	MCMC setting (default = 0.95).
`max.treedepth`	MCMC setting (default = 12).

The input to IgGeneUsage is a table with usage frequencies for each gene of a repertoire that belongs to a particular biological condition. For the analysis of differential gene usage between two biological conditions, IgGeneUsage employs a Bayesian hierarchical model for zero-inflated beta-binomial (ZIBB) regression (see vignette 'User Manual: IgGeneUsage').

`glm.summary`	differential gene usage statistics for each gene. 1) es = effect size on differential gene usage (mean, median standard error (se), standard deviation (sd), L (low boundary of HDI), H (high boundary of HDI); 2) contrast = direction of the effect; 3) pmax = probability of differential gene usage
`test.summary`	differential gene usage statistics computed with the Welch's t-test (columns start with 't'), and Wilcoxon signed-rank test (columns start with 'u'). For both test report P-values, FDR-corrected P-values, Bonferroni-corrected P-values. Additionally, we report t-value and 95% CI (from the t-test) and U-value (from the Wilcoxon signed-rank test).
`glm`	stanfit object
`ppc.data`	two types of posterior predictive checks: 1) repertoire- specific, 2) gene-specific
`usage.data`	processed gene usage data used for the model

Simo Kitanovski <simo.kitanovski@uni-due.de>

LOO, Ig, IGHV_Epitopes, IGHV_HCV, Ig_SE

## Not run:

# input data
data(Ig)
head(Ig)

# Alternative:
# use SummarizedExperiment input data
# data(Ig_SE)


# run differential gene usage (DGU)
M <- DGU(usage.data = Ig,
         mcmc.warmup = 250,
         mcmc.steps = 1000,
         mcmc.chains = 2,
         mcmc.cores = 1,
         hdi.level = 0.95,
         adapt.delta = 0.95,
         max.treedepth = 13)

## End(Not run)