RTP

RTP computes the significance of rank truncated products of P-values to combine evidence from subsets of hypotheses in multiple-testing contexts such as genome-wide association scans.

Key Features:

• Rank Truncated Product: Computes the product of the K most significant P-values, where K is a predetermined number often chosen based on a hypothesized disease model.
• Independence from Sample Size: The selection of K is independent of sample size, maintaining the test's applicability across different study scales.
• Fixed Effects Hypothesis Alignment: Focuses on top-ranking P-values to align with experimental scenarios in which loci have fixed effects.
• Increased Power and Quality: Enhances power to detect genome-wide associations by identifying a manageable candidate set of significant hypotheses for follow-up.
• Accounting for Correlated Tests: Includes methods to account for correlation among tests common in genomewide scans.

Scientific Applications:

• Genomewide Association Scans: Combines evidence across top P-values to prioritize variants associated with traits or diseases.
• Exploratory Studies with Multiple Hypotheses: Prioritizes a subset of promising hypotheses in large-scale exploratory analyses involving many tests.

Methodology:

Forms the product of the K most significant P-values and evaluates its significance using the distribution of the rank truncated product, incorporating procedures to account for correlated tests and control multiple testing.