RSC-based differential model with correlation removal for improving multi-omics clustering.

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Multi-omics data integration is vital for accurate cancer subtyping.
• Correlations within and between omics datasets can introduce noise and reduce clustering performance.
• Eliminating redundant correlational information is crucial for robust multi-omics analysis.

Purpose of the Study:

• To develop a novel method for improving multi-omics clustering by addressing data correlations.
• To introduce the RSC-based differential model with correlation removal (RSC-MCR) for enhanced cancer subtyping.
• To evaluate the efficacy of RSC-MCR compared to existing decorrelation techniques.

Main Methods:

• Utilized Relative-Skill-Correlation (RSC) to compute pairwise feature correlations across different omics layers.
• Developed a differential model to quantify and remove redundant correlations between omics data.
• Integrated the RSC-MCR method with various clustering algorithms (CC, FCM, SNF, NMF, LRAcluster).

Main Results:

• RSC-MCR effectively identified and removed redundant correlations present in multi-omics datasets.
• The proposed method demonstrated significant improvements in clustering performance across multiple cancer datasets.
• Comparative analyses confirmed the superiority of RSC-MCR over other decorrelation strategies.

Conclusions:

• RSC-MCR offers an effective strategy for enhancing multi-omics clustering by managing data correlations.
• The method holds promise for more accurate cancer subtyping and biological discovery.
• This work provides a valuable tool for researchers working with complex multi-omics data.