IMIX: a multivariate mixture model approach to association analysis through multi-omics data integration.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Integrative genomic analysis leverages multiplatform high-dimensional data (DNA methylation, copy number variation, gene expression) to study complex diseases.
• Current methods often analyze data types separately, leading to reduced statistical power and uncontrolled false discovery rates (FDR).

Purpose of the Study:

• To develop a multivariate mixture model (IMIX) for integrating multiple genomic data types.
• To improve statistical power and control the overall false discovery rate (FDR) in integrative analyses.
• To provide novel multi-omics insights into complex diseases.

Main Methods:

• Proposed a multivariate mixture model (IMIX) framework to integrate diverse genomic data types.
• Modeled inter-data-type correlations within the IMIX framework.
• Investigated across-data-type FDR control and compared performance against existing methods via simulations.

Main Results:

• IMIX demonstrated lower misclassification rates at controlled overall FDR compared to individual data type analysis strategies.
• Simulations confirmed IMIX's superior performance over Benjamini-Hochberg FDR control, q-value, and local FDR control.
• Applied IMIX to The Cancer Genome Atlas data, revealing new multi-omics insights into bladder cancer subtypes and pancreatic cancer prognosis.

Conclusions:

• IMIX offers a statistically principled, computationally efficient framework for integrative genomic analysis.
• The IMIX method effectively controls FDR and enhances statistical power when analyzing multi-omics data.
• IMIX provides valuable insights into the genetic architecture and mechanisms of complex diseases.