PaM

PaM optimizes treatment and case-control pairings using demographic and genetic data to minimize population stratification bias and increase genetic homogeneity in clinical trials and genetic studies.

Key Features:

• Optimization of Pairing Assignments: Performs pairing optimization both a priori and a posteriori to generate matched treatment-control pairs.
• Dual-criteria Integration: Integrates demographic and genetic data as joint criteria for matching decisions.
• Reduction of Population Stratification Bias: Minimizes bias arising from uneven ancestry distributions between treatment and control groups.
• Improved Genetic Homogeneity Compared to PCA: Produces matched pairs that are more genetically homogeneous than those identified using principal component analysis (PCA).
• Validation on Simulated and Real Datasets: Matching performance has been evaluated using both simulated datasets and empirical real-world datasets.
• Ancestry Inference for Precision Medicine: Infers ancestry information to support identification of responders and precision medicine analyses.

Scientific Applications:

• Clinical Trial Matching: Optimizes treatment-control matching in clinical trials from pilot phases through Phase-III to improve assessment of drug efficacy and reproducibility.
• Precision Medicine: Enables characterization of responders within trials using ancestry-informed matching to support precision medicine strategies.
• Genetic Association Studies: Provides precise case-control matching to reduce confounding from ancestry in genetic studies.

Methodology:

Uses a model-based approach that integrates demographic and genetic data for dual-criteria optimization, supports both a priori and a posteriori pairing, and has been validated on simulated and real datasets with comparisons to principal component analysis (PCA).