ManiNetCluster

ManiNetCluster identifies and aligns local and non-linear structures in gene networks to reveal cross-network functional links across conditions such as time, disease, organism, and environmental perturbations.

Key Features:

• Manifold Learning Approach: Leverages manifold learning to uncover and align local and non-linear structures within gene networks.
• Simultaneous Alignment and Clustering: Performs simultaneous alignment and clustering of gene networks, including co-expression networks, to systematically explore functional connections across conditions.
• Cross-Network Functional Link Identification: Identifies cross-network functional links to reveal how genes interact across different biological contexts.

Scientific Applications:

• Orthologous Gene Alignment: Demonstrated superior alignment of orthologous genes based on developmental expression profiles across model organisms (p-value < 2.2×10^-16).
• Time Series Transcriptome Analysis: Applied to time series transcriptome data from Chlamydomonas reinhardtii to identify genomic functions linking metabolic processes between light and dark periods in a diurnally cycling culture and to pinpoint genes potentially regulating these processes.

Methodology:

ManiNetCluster employs manifold learning to detect and match non-linear local and global structures among gene networks and performs simultaneous alignment and clustering of networks (e.g., co-expression networks).