MSClust

MSClust clusters 16S rRNA sequences into operational taxonomic units (OTUs) for microbiome profiling of next-generation sequencing data and is implemented as a MATLAB package.

Key Features:

• Multi-Seed Based Heuristic Clustering: Uses multiple seeds to form candidate clusters, improving robustness of OTU inference compared with single-seed approaches.
• Greedy Network Clustering Strategy: Processes reads with a greedy network clustering strategy to enable rapid handling of large sequence datasets.
• Balanced Accuracy and Efficiency: Optimizes inference accuracy while reducing memory usage relative to existing heuristic methods.
• Scalability and Biological Relevance: Demonstrated superior biological accuracy and scalability over other heuristic clustering methods in numerical examples.

Scientific Applications:

• Microbiome profiling: Clustering 16S rRNA reads into OTUs for characterization of microbial community composition.
• Ecological studies: Enables analysis of microbial diversity and community structure in environmental samples.
• Health-related microbiome research: Supports studies of human and animal microbiomes linking OTU-level composition to health or disease.
• Environmental monitoring: Facilitates monitoring of microbial changes across environmental gradients or pollution studies.

Methodology:

Implements multi-seed based heuristic clustering and a greedy network clustering strategy; implemented as a MATLAB package and reported to use less memory than other heuristic methods.