sponge_compute_p_values: Compute p-values for SPONGE interactions

In SPONGE: Sparse Partial Correlations On Gene Expression

Description

This method uses pre-computed covariance matrices that were created for various gene-gene correlations (0.2 to 0.9 in steps of 0.1) and number of miRNAs (between 1 and 8) under the null hypothesis that the sensitivity correlation is zero. Datasets are sampled from this null model and allow for an empirical p-value to be computed that is only significant if the sensitivity correlation is higher than can be expected by chance given the number of samples, correlation and number of miRNAs. p-values are adjusted indepdenently for each parameter combination using Benjamini-Hochberg FDR correction.

Usage

sponge_compute_p_values(sponge_result, null_model, log.level = "ERROR")

Arguments

sponge_result	A data frame from a sponge call
null_model	optional, pre-computed simulated data
log.level	The log level of the logging package

Value

A data frame with sponge results, now including p-values and adjusted p-value

See Also

sponge_build_null_model

Examples

sponge_compute_p_values(ceRNA_interactions,
null_model = precomputed_null_model)

SPONGE documentation built on Nov. 8, 2020, 5:39 p.m.