Clustering single-cell multi-omics data via graph regularized multi-view ensemble learning.

Area of Science:

• Computational Biology
• Genomics
• Bioinformatics

Background:

• Single-cell clustering is vital for cell type identification and analyzing cellular heterogeneity.
• Existing methods using only gene expression (mono-omic) data often yield suboptimal results due to limited information.
• Integrating multi-omics data offers richer insights but presents challenges in effective data fusion.

Purpose of the Study:

• To develop an advanced single-cell clustering model capable of adaptively integrating diverse multi-omics data.
• To address the challenge of effectively combining information from multiple omics sources for improved cell clustering.
• To create a robust method that performs well across various types of single-cell multi-omics datasets.

Main Methods:

• Proposes a graph-regularized multi-view ensemble clustering (GRMEC-SC) model.
• The model adaptively integrates multiple omics data sources.
• Leverages insights from multiple base clustering results for enhanced performance.

Main Results:

• The GRMEC-SC model demonstrates competitive performance on five diverse multi-omics datasets.
• Evaluations confirm the model's effectiveness across datasets with varying characteristics.
• The approach successfully integrates multiple omics data for improved single-cell clustering.

Conclusions:

• The GRMEC-SC model provides an effective solution for single-cell clustering using multi-omics data.
• The method shows robust performance and adaptability to different data types.
• This work advances the integration of multi-omics data for biological discovery.