SweepCluster

SweepCluster identifies gene-specific selective sweeps in microbial populations by clustering SNPs and estimating significance using an anchor-extension method to detect regions of concentrated polymorphisms within genes.

Key Features:

• Gene-Centric Approach: Focuses on identifying gene regions with high spatial clustering of polymorphisms pre-selected based on elevated population subdivision, excessive linkage disequilibrium, or significant phenotype association.
• Spatial-Aware Analysis: Incorporates the spatial distribution of mutations within gene regions to improve resolution and accuracy of sweep detection.
• Performance and Sensitivity: Simulation studies reported accuracy and sensitivity exceeding 90% for its clustering algorithms.
• Application Versatility: Applicable to prokaryotic and eukaryotic genotype datasets, including human genotype data, and capable of recovering known sweep regions under varied pre-selection parameters.
• Reduction of Uninformative Signals: Polymorphism pre-selection substantially reduces uninformative signals in bacterial datasets such as Streptococcus pyogenes and Streptococcus suis.
• Validation and Concordance: Validation on Vibrio cyclitrophicus genotype data yielded a concordance rate of 78%, with potential underestimation attributed to reference genome and clustering strategy differences.

Scientific Applications:

• Adaptive Evolution: Identify gene-specific selective sweeps to investigate adaptive evolution in microbial populations.
• Ecological Differentiation: Detect sweeps associated with ecological differentiation among microbial lineages.
• Phenotypic Divergence: Associate gene-specific sweeps with phenotypic traits via pre-selection by phenotype association.
• Cross-Domain Genomic Analysis: Analyze selective sweeps across prokaryotic and eukaryotic genotype datasets to compare sweep patterns.

Methodology:

SNP clustering and significance estimation are performed using an anchor-extension method with polymorphism pre-selection based on population subdivision, linkage disequilibrium, or phenotype association, combined with spatial-aware analysis of mutation distributions; performance was evaluated by simulation studies and validation on genotype datasets (e.g., Vibrio cyclitrophicus, Streptococcus pyogenes, Streptococcus suis, and human genotype data).