AcaFinder

AcaFinder identifies anti-CRISPR associated (aca) genes and acr-aca operons in prokaryotic and phage genomes to enable discovery and analysis of Acr (Anti-CRISPR)–Aca interactions and regulation of CRISPR-Cas systems.

Key Features:

• Guilt-by-association (GBA) prediction: Uses a guilt-by-association strategy to predict aca genes and acr-aca operonic structures.
• HMM-based homolog detection: Employs Hidden Markov Models (HMMs) to identify homologs of known Aca proteins.
• Input scope: Accepts genomic inputs including prophages, CRISPR-Cas loci, and self-targeting spacers (STSs).
• Large-scale screening: Applied to over 16,000 prokaryotic genomes and 142,000 gut phage genomes for genome-wide mining.
• Multistep filtering and high-confidence families: Uses multistep filtering to identify 36 new high-confidence Aca families.
• Taxon-specific discovery: Detected seven novel Aca families from Bacteroidota, Actinobacteria, and Fusobacteria and expanded known diversity beyond Proteobacteria and Firmicutes.
• Operonic colocalization network analysis: Analyzes operonic colocalizations to reveal association networks and modular acr-aca operon combinations.
• HTH domain and adjacency criterion: Leverages the conserved helix-turn-helix (HTH) domain and adjacency to Acr homologs within the same operon as indicators of Aca candidates.

Scientific Applications:

• Genome mining: Enables systematic discovery of aca genes and acr-aca operons across large prokaryotic and phage genome datasets.
• Discovery of new Aca families: Supports identification of novel Aca families, including seven from human gut bacteria and 36 high-confidence families overall.
• Operon evolution and modularity studies: Facilitates analysis of recombination and modular assembly of acr-aca operons via colocalization networks.
• CRISPR-Cas regulation research: Provides candidate Acas and hypotheses for studying regulation of CRISPR-Cas systems and interactions between Acr proteins and Acas.

Methodology:

Computational methods explicitly include guilt-by-association prediction, Hidden Markov Model (HMM) searches for Aca homologs, multistep filtering of candidates, and operonic colocalization network analysis.