DGU: Differential gene usage in immune repertoires

In IgGeneUsage: Differential gene usage in immune repertoires

Description

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

Usage

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

Arguments

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).

Details

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').

Value

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

Author(s)

Simo Kitanovski <[e-mail]>

See Also

LOO, Ig, IGHV_Epitopes, IGHV_HCV, Ig_SE

Examples

# 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 = 500,
         mcmc.steps = 1500,
         mcmc.chains = 2,
         mcmc.cores = 1,
         hdi.level = 0.95,
         adapt.delta = 0.95,
         max.treedepth = 13)

IgGeneUsage documentation built on Nov. 8, 2020, 7:47 p.m.