Reciprocal Best Hits - Best Bidirectional Hits

Reciprocal Best Hits - Best Bidirectional Hits identifies reciprocal best-scoring protein matches between two genomes to infer orthology and enable prediction of functional gene couplings.

Key Features:

• Reciprocal best-hit identification: Detects pairs of proteins from two genomes that are each other's highest-scoring match in the opposite genome.
• Orthology proxy: Uses RBH as a widely accepted proxy for orthology based on evolutionary conservation of protein sequences.
• Primary aligner: Performs sequence comparisons using NCBI BLAST as the standard method for detecting homologs and RBHs.
• Alternative aligners evaluated: Includes comparisons to BLAT, LAST, and UBLAST as reported in comparative studies (PMID: 25013894).
• Performance characteristics: Alternative programs reduced computation time relative to BLAST (ranging from a hundredth to a 25th of BLAST time), with LAST closely matching BLAST for homolog and RBH detection and UBLAST yielding 0.6–0.8 of the RBHs detected by BLAST between dissimilar genomes.
• Limitations with BLAT: BLAT is optimized for very similar sequences and missed many homologs and RBHs in more divergent genomes.
• Functional coupling prediction: Extends RBH use to detect conserved gene clusters and predict functional coupling between genes, validated in prokaryotic genome analyses (PMID: 10077608).
• Empirical dataset scale: Demonstrated on analyses of over 30 complete or nearly complete genomes for detection of functionally coupled gene classes such as transport and signal transduction clusters.
• Computational cost note: Sequence comparison can be computationally intensive and time-consuming, motivating use of faster aligners.

Scientific Applications:

• Ortholog detection: Identification of putative orthologous protein pairs between genomes for comparative genomics and evolutionary analysis.
• Homolog detection across divergence: Detection and comparison of homologs across genomes of varying similarity using BLAST, LAST, BLAT, or UBLAST.
• Functional coupling prediction: Prediction of functionally coupled genes via conservation of gene clusters across prokaryotic genomes, supporting assignment of functions to uncharacterized genes.
• Genome annotation support: Use of RBH relationships to inform annotation and functional inference in newly sequenced genomes.

Methodology:

Compute pairwise protein sequence comparisons between two genomes (using NCBI BLAST or alternatives BLAT, LAST, UBLAST), identify each protein's highest-scoring match in the opposite genome, and retain pairs that are reciprocal best-scoring matches as RBHs.