HipMCL

HipMCL performs parallel Markov clustering of large-scale biological networks to identify highly connected regions for analysis of gene expression and protein-protein interaction networks.

Key Features:

• Parallel Implementation: A parallelized version of the Markov Clustering (MCL) algorithm optimized for distributed-memory computing environments.
• Scalability: Can leverage up to 2000 compute nodes to handle networks of approximately 70 million nodes and 68 billion edges.
• Efficiency: Demonstrated ability to cluster massive networks in roughly 2.4 hours, offering performance improvements of several orders of magnitude over traditional MCL implementations.
• Technology Stack: Implemented using MPI (Message Passing Interface) and OpenMP (Open Multi-Processing) for parallel processing and resource utilization.

Scientific Applications:

• Large-scale network clustering: Identification of highly connected regions in biological networks such as protein-protein interaction and gene expression networks.
• Functional module detection: Discovery of groups of genes, proteins, or molecular entities with shared functional affinities or structural similarities.
• Gene expression and PPI analysis: Support for analysis of gene expression patterns and protein-protein interactions at scales that exceed standard MCL capabilities.

Methodology:

Parallelized Markov Clustering (MCL) implemented with MPI and OpenMP, distributing computational tasks across multiple nodes in a distributed-memory environment.