AMEND 2.0: module identification and multi-omic data integration with multiplex-heterogeneous graphs.

Area of Science:

• Systems biology
• Bioinformatics
• Network analysis

Background:

• Multi-omic studies offer comprehensive biological insights by examining cellular changes.
• Biological networks are effective for omics analysis but current methods lack generalizability for diverse data types.
• Existing network methods are often limited to specific omics types, hindering broad application.

Purpose of the Study:

• To present AMEND 2.0, an advanced method for active module identification.
• To enable analysis of multiplex and heterogeneous networks integrated with multi-omic data.
• To provide a highly generalizable framework for diverse biological network analyses.

Main Methods:

• AMEND 2.0 utilizes Random Walk with Restart for multiplex-heterogeneous networks.
• Incorporates degree bias adjustment and biased random walk for multi-objective module identification.
• Applied to renal cell carcinoma and O-GlcNAc Transferase knockout multi-omic datasets.

Main Results:

• AMEND 2.0 demonstrates generalizability across multiplex and heterogeneous networks.
• The method successfully analyzed diverse multi-omic datasets.
• Evaluated performance of subroutines for node ranking and degree bias adjustment.

Conclusions:

• Network-based multi-omic data analysis requires advanced methods for deeper biological understanding.
• AMEND 2.0 integrates network analysis techniques into a versatile tool for diverse scenarios.
• The software is available in R for broader accessibility and application.