Randomized singular value decomposition for integrative subtype analysis of 'omics data' using non-negative matrix...

Area of Science:

• Bioinformatics
• Computational Biology
• Cancer Research

Background:

• Multi-omics data integration is crucial for cancer subtype differentiation.
• Matrix factorization methods like Non-negative Matrix Factorization (NMF) are used for integrative clustering.
• High dimensionality and long computation times are challenges in multi-omics clustering.

Purpose of the Study:

• To propose a novel method, intNMF-rsvd, for efficient integrative clustering of multi-omics data.
• To address the challenges of high dimensionality and computational cost in cancer subtype analysis.
• To improve the identification of latent subtype structures across diverse omics datasets.

Main Methods:

• Utilized randomized singular value decomposition (RSVD) for dimensionality reduction.
• Applied Non-negative Matrix Factorization (NMF) for integrative clustering (intNMF-rsvd).
• Projected multi-omics data into an eigen vector space with a user-specified lower rank.

Main Results:

• intNMF-rsvd demonstrated efficient and competitive performance compared to state-of-the-art methods.
• The method effectively handles large numbers of features and significantly reduces computation time.
• Evaluated using simulated and The Cancer Genome Atlas (TCGA) datasets.

Conclusions:

• intNMF-rsvd offers an efficient approach for multi-omics integrative clustering.
• The method's ability to reduce computation time makes it suitable for large-scale cancer data analysis.
• Identified subtypes can facilitate further clinical association studies for disease etiology.