PTest

PTest implements a partition test to analyze multiple SNP p-values simultaneously for detecting genetic associations in case-control studies and estimating the number of functional SNPs underlying complex traits.

Key Features:

• Partition Test Methodology: Employs a partition test to identify the smallest subset of p-values that deviate most from the expected random (uniform) distribution on the interval 0 to 1.
• Handling Multiple Variants: Aggregates signals from multiple genetic variants, each potentially of small effect, rather than relying on single-SNP analyses.
• Power and Significance: Includes power calculations showing increased ability to detect associations compared with SNP-by-SNP analyses and procedures estimating the false discovery rate.
• Estimation of Functional SNPs: Produces an estimate of the number of functional SNPs contributing to a trait or disease.
• Validation on Published Datasets: Applied to six published datasets where it outperformed well-known procedures in identifying significant associations.

Scientific Applications:

• Case-control genetic association studies: Detects associations by jointly analyzing SNP p-values derived from genotype or allele frequency comparisons between cases and controls.
• Genome-wide association studies (GWAS): Applicable to GWAS for identifying polygenic signals across many SNPs.
• Analysis of polygenic complex traits: Suitable for studying traits influenced by multiple small-effect variants to infer genetic architecture.
• Human gene mapping: Supports human gene mapping by aggregating multi-variant effects and estimating numbers of functional SNPs.

Methodology:

Performs partition testing on p-values (0–1) derived from single-SNP tests such as chi-square comparisons of genotype or allele frequencies between cases and controls; aggregates multiple-variant effects; conducts power calculations; estimates the number of functional SNPs; validated on six published datasets.