DBSCAN-SWA

DBSCAN-SWA predicts prophage regions in bacterial genomes to identify integrated bacteriophages and support analyses of horizontal gene transfer and phage–host interactions.

Key Features:

• Speed: Operates faster than previous prophage prediction tools, enabling large-scale genomic analyses.
• Detection Power: Achieves 85% recall on raw DNA sequences against a benchmark of 184 manually curated prophages, compared with Phage_Finder (63%), VirSorter (74%), and PHASTER (82%).
• Algorithm: Combines DBSCAN (Density-Based Spatial Clustering of Applications with Noise) with a sliding window approach (SWA) to identify prophage regions.
• Input Data: Processes raw DNA sequences for prophage prediction.
• Implementation: Implemented in Python3.

Scientific Applications:

• Phage Therapy: Facilitates characterization of prophages to inform phage–host dynamics relevant to phage therapy and studies of antibiotic resistance.
• Microbial Ecology: Enables investigation of prophage roles in bacterial evolution and horizontal gene transfer within microbial communities.
• Genomic Research: Supports precise annotation of bacterial genomes for diverse genomic investigations.

Methodology:

Applies DBSCAN clustering combined with a sliding window approach (SWA) to detect density-defined genomic regions indicative of prophages from raw DNA sequences.