NESM: a network embedding method for tumor stratification by integrating multi-omics data.

Area of Science:

• Computational biology
• Genomics
• Bioinformatics

Background:

• Tumor stratification is crucial for cancer diagnosis and personalized treatment.
• High-throughput sequencing generates vast multi-omics data, enabling advanced cancer subtyping.
• Integrating diverse molecular data offers a comprehensive view of cancer heterogeneity.

Purpose of the Study:

• To develop and validate a Network Embedding method for tumor stratification by integrating multi-omics data.
• To leverage gene features, somatic mutations, and network topology for improved cancer classification.
• To identify novel cancer subtypes associated with patient survival outcomes.

Main Methods:

• Network Embedding for tumor Stratification (NETS) method was developed.
• NETS integrates DNA methylation, mRNA expression, and protein-protein interaction data.
• Supervised (Light Gradient Boosting Machine) and unsupervised (DBSCAN) learning algorithms were employed.

Main Results:

• The NETS method achieved the highest Area Under the Curve (AUC) for cancer type stratification compared to three other methods, with an average AUC of 0.91.
• Extracted patient features using NETS proved effective for tumor stratification.
• Unsupervised clustering identified cancer subtypes significantly associated with patient survival.

Conclusions:

• The Network Embedding method for tumor Stratification by integrating Multi-omics is a powerful tool for cancer classification.
• NETS effectively utilizes multi-omics data to reveal cancer heterogeneity and identify clinically relevant subtypes.
• This approach holds promise for advancing precision oncology and improving patient outcomes.