MVDA: a multi-view genomic data integration methodology.

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• High-throughput omics technologies generate molecular profiling data for individuals.
• Integrating multi-modal omics data enhances patient subclassification power compared to separate analyses.

Purpose of the Study:

• To develop and evaluate a multi-view approach for integrating omics data.
• To improve the identification of clinically relevant patient subgroups.

Main Methods:

• A late integration strategy factorizes membership matrices from single-view clustering.
• The method was evaluated on six multi-view cancer datasets.

Main Results:

• Statistically significant patient subclasses were identified across all datasets.
• Novel subgroups were discovered, and the method outperformed existing multi-view clustering algorithms.
• The approach quantifies the contribution of individual data views to the final results.

Conclusions:

• Integrating prior information with genomic features is effective for disease subgroup identification.
• The developed methodology is available as open-source R code.