dks

dks evaluates the validity of p-values from multiple testing procedures in genomics by analyzing simulated datasets to detect anomalies in p-value distributions.

Key Features:

• Diagnostic assessment of p-values: Provides a suite of functions to assess the accuracy of p-values generated by multiple testing procedures.
• Simulated-data input: Operates on complete sets of p-values derived from multiple simulated studies.
• Joint distribution evaluation: Evaluates the joint distribution of p-values across simulations to assess correctness of testing methods.
• Spectrum-wide analysis: Analyzes the entire spectrum of p-values across simulations to detect anomalies or inconsistencies.
• Focus on high-throughput data: Targets validation of multiple testing procedures used in high-throughput genomic analyses.
• Implementation platform: Distributed as a package within the Bioconductor project for use in R.

Scientific Applications:

• Validation of multiple testing methods: Verifies whether p-values produced by multiple testing procedures are statistically valid.
• Simulation-based method development: Enables testing and refinement of statistical approaches using controlled simulated datasets before application to real biological data.
• Assessment of reproducibility and robustness: Supports evaluation of robustness and reproducibility of inference in genomics driven by multiple testing.

Methodology:

Uses simulated datasets comprising complete sets of p-values from multiple studies, evaluates the joint distribution of p-values, and analyzes the full p-value spectrum across simulations to detect anomalies or inconsistencies.