Bootstrap Evaluation of Association Matrices (BEAM) for Integrating Multiple Omics Profiles with Multiple Outcomes

Area of Science:

• Genomics
• Biostatistics
• Bioinformatics

Background:

• Integrating multi-omics data with clinical endpoints is crucial for scientific discovery.
• Existing statistical methods struggle to accommodate diverse omics profiles and clinical outcomes simultaneously.

Purpose of the Study:

• To introduce Bootstrap Evaluation of Association Matrices (BEAM), a novel statistical method for integrating multiple omics profiles with multiple clinical endpoints.
• To provide a flexible and robust tool for identifying significant gene-endpoint associations.

Main Methods:

• BEAM utilizes regression models to associate sets of omic features with clinical endpoints.
• Bootstrap resampling is employed to determine the statistical significance of these associations.
• The method is designed to handle an arbitrary number of omics profiles and clinical endpoints.

Main Results:

• Simulations show BEAM performs comparably to optimal tests and surpasses other integrated analysis methods.
• In a pediatric leukemia study, BEAM identified biologically relevant genes missed by univariate and other integrated approaches.
• BEAM demonstrated superior performance in identifying key genes compared to existing methods.

Conclusions:

• BEAM is a powerful, flexible, and robust tool for multi-omics data integration.
• The method facilitates the identification of genes for further laboratory and clinical research.
• BEAM offers a significant advancement in analyzing complex biological datasets.