Multi-omics Integrative Analysis for Incomplete Data Using Weighted p-Value Adjustment Approaches

Area of Science:

• Bioinformatics
• Genomics
• Statistical Genetics

Background:

• High-throughput technologies enable multi-omics data acquisition from single individuals.
• Missing values are common in multi-omics data due to invasive sampling, complicating joint analyses.
• Existing methods like complete case analysis or multiple imputation have limitations.

Purpose of the Study:

• To propose a novel integrative multi-omics analytical framework.
• To address the challenge of missing values in joint multi-omics data analysis.
• To enhance statistical power by incorporating incomplete datasets.

Main Methods:

• Developed a framework based on p-value weight adjustment.
• Split data into complete and incomplete sets.
• Derived weights and weight-adjusted p-values to integrate all observations.

Main Results:

• Simulation analyses demonstrated considerable statistical power gains.
• The proposed framework outperformed complete case analysis and multiple imputation.
• Successfully applied to a preterm infant birth weight study involving DNA methylation and mRNA data.

Conclusions:

• The p-value weight adjustment framework effectively incorporates incomplete multi-omics data.
• Offers a powerful alternative for joint analysis of multi-omics datasets.
• Facilitates more comprehensive insights in biomedical studies with missing data.