2SigFinder

2SigFinder predicts genomic islands in prokaryotic genomes using a multiscale statistical algorithm that detects deviations in genomic composition across multiple window sizes.

Key Features:

• Multiscale Statistical Detection: Integrates small-scale and large-scale statistical testing to identify genomic regions with atypical genomic signatures.
• Local Deviation Analysis: Applies small-scale tests to detect local deviations from host genome composition using large-scale genomic features.
• Segment-Level Identification: Uses large-scale statistical testing with small-scale features to identify multi-window genomic segments corresponding to genomic islands.
• Annotation-Independent Detection: Predicts genomic islands without requiring annotated genome information or prior reference datasets.

Scientific Applications:

• Genomic Island Identification: Detects horizontally acquired genomic regions in prokaryotic genomes.
• Microbial Pathogenicity Studies: Supports identification of pathogenicity islands and virulence-associated genomic regions.
• Mobile Genetic Element Analysis: Enables detection of genomic island-associated features such as tRNA genes, phage regions, and homing endonuclease genes.

Methodology:

The algorithm performs small-scale statistical tests to detect local compositional deviations using large-scale genomic features, followed by large-scale statistical testing with small-scale features to identify multi-window genomic segments corresponding to genomic islands.