ALFATClust

ALFATClust clusters biological sequences using rapid alignment-free pairwise distance calculations and graph-based community detection to produce robust clusters with dynamically determined per-cluster cut-off thresholds for genomic, proteomic, and metagenomic analyses.

Key Features:

• Alignment-Free Sequence Distance Calculation: ALFATClust computes rapid alignment-free pairwise sequence distances to reduce computational overhead compared with alignment-based approaches.
• Dynamic Thresholding: It determines cut-off thresholds individually for each cluster based on intra-cluster similarity and inter-cluster separation.
• Community Detection in Graphs: The tool employs community detection algorithms on a sequence similarity graph to generate clusters.
• Benchmarking Superiority: Comparative analyses demonstrate ALFATClust generally outperforms existing clustering approaches by maintaining robust cluster formation and substantial separation between clusters across benchmark datasets.
• Quality Evaluation Report: ALFATClust provides evaluation reports to verify the quality of non-singleton clusters, reporting intra-cluster similarity and inter-cluster differentiation.

Scientific Applications:

• Genomic Research: Facilitates clustering of genomic sequences for studies in evolutionary biology, population genetics, and comparative genomics.
• Proteomics: Groups protein sequences to identify functional similarities or evolutionary relationships.
• Metagenomics: Supports analysis of complex microbial community sequence data by efficiently clustering diverse sequences.

Methodology:

ALFATClust computes alignment-free pairwise sequence distances, constructs a sequence similarity graph, applies community detection algorithms, dynamically determines per-cluster cut-off thresholds based on intra- and inter-cluster measures, and generates evaluation reports for non-singleton clusters.