Rephine.r

Rephine.r refines gene calls and gene clusters to improve bacteriophage pangenome construction and phylogenetic inference.

Key Features:

• Error Correction: Identifies and corrects three primary sources of fragmented gene calls: indels causing early stop codons and new start codons, interruptions by selfish genetic elements (e.g., homing endonucleases or intron-like sequences), and splitting at the ends of reported genomes.
• Gene Fusion: Fuses identified fragmented genes to reconstruct coherent sequences and produce accurate multiple sequence alignments.
• Homolog Identification: Uses Hidden Markov Models (HMMs) to search for distant homologs and merge gene families that were erroneously separated.
• Iterative Refinement: Performs a final round of fragment identification to further refine gene clusters.
• Outcome Improvement: Recovers additional single-copy core genes and increases bootstrap support in inferred phylogenetic trees.

Scientific Applications:

• Bacteriophage pangenomics: Refines gene clustering in bacteriophage genomes, including recovery of single-copy core genes affected by homing endonucleases or intron-like sequences.
• Phylogenetic inference: Improves robustness of phylogenetic trees and bootstrap support by providing corrected gene alignments and clusters.
• Comparative genomics of specific phage groups: Applied to Tevenvirinae (e.g., T4), Studiervirinae (e.g., T7), and Pbunaviruses (e.g., PB1) to recover additional core genome members and enhance bootstrap support.

Methodology:

The pipeline first identifies fragmented gene calls and fuses them into coherent sequences and alignments, then uses Hidden Markov Models (HMMs) to search for distant homologs and merge separated gene families, followed by a final iterative round of fragment identification to refine clusters.