SpydrPick

SpydrPick identifies genome-wide co-evolving loci by calculating Mutual Information and pruning indirect couplings to detect epistasis and co-selection in population genomics datasets.

Key Features:

• Direct Coupling Analysis: Performs direct coupling analysis on aligned categorical datasets by calculating Mutual Information (MI) scores between node pairs and applying thresholds to construct a coupling graph.
• Indirect Coupling Pruning: Applies the ARACNE algorithm to remove edges representing indirect couplings, retaining direct interactions.
• Model-Free Population-Structure Correction: Incorporates an efficient correction for population structure to adjust for phylogenetic signals without requiring a predefined phylogenetic tree.
• Scalability: Processes large-scale datasets, including bacterial pan-genomes, for genome-wide analyses of epistasis and co-selection.
• Visualization: Produces Manhattan-like plots to display genome-wide signals of co-evolution and highlight significant couplings.

Scientific Applications:

• Detection of co-selection and epistasis: Identifies dependencies between polymorphisms indicating co-selective pressures or epistatic interactions in bacterial populations by analyzing population-level covariation of alleles across chromosomes.
• Identification of loci linked to virulence and antibiotic resistance: Has been applied to datasets from Streptococcus pneumoniae and Neisseria meningitidis to recover known and novel targets of co-selection related to virulence and antibiotic resistance.
• Pan-genome and comparative population genomics: Enables genome-wide interaction analyses across multiple bacterial species and large population datasets.

Methodology:

Calculates Mutual Information between aligned categorical loci, applies thresholds to build a coupling graph via direct coupling analysis, prunes indirect edges using ARACNE, and applies a model-free correction for population structure without requiring a predefined phylogenetic tree.