HHblits

HHblits performs sensitive HMM-HMM iterative sequence searches to detect remote protein homologs and support protein function and structure prediction.

Key Features:

• Profile HMM Representation: Represents query sequences and database entries as profile Hidden Markov Models (HMMs) to capture evolutionary information of protein families.
• Iterative Search Methodology: Performs iterative HMM-HMM comparisons that refine alignments across successive search rounds to increase sensitivity.
• Prefiltering Mechanism: Uses a discretized-profile prefilter to rapidly narrow candidate matches before full HMM-HMM alignment, accelerating searches.
• Enhanced Sensitivity and Accuracy: Demonstrates approximately 50–100% higher sensitivity than PSI-BLAST in benchmarks, improving detection of distant homologs relative to PSI-BLAST and HMMER3.

Scientific Applications:

• Protein Function Prediction: Aligns queries to annotated proteins to support prediction of protein function.
• Structural Biology: Enables inference of structural features and identification of templates for modeling protein structures from sequence similarity.
• Evolutionary Studies: Identifies distant homologs to characterize evolutionary relationships and protein family diversity.

Methodology:

Represents sequences as profile HMMs; performs iterative HMM-HMM comparisons with refinement across successive rounds; applies a discretized-profile prefilter to accelerate candidate selection.