FastHMM

FastHMM accelerates identification of protein family members in large protein sequence databases by combining PSI-BLAST heuristics with HMMer profile HMM confirmation.

Key Features:

• PSI-BLAST integration: FastHMM uses Position-Specific Iterated BLAST (PSI-BLAST) to rapidly identify candidate sequences related to a protein family.
• HMMer confirmation: Candidate hits are confirmed using HMMer profile hidden Markov model (HMM) searches to improve specificity.
• Heuristic alternative: Provides a fast heuristic alternative to traditional HMM search, InterProScan, and all-versus-all BLAST for large-scale sequence analysis.
• FastBLAST and rpsblast integration: Integrates with FastBLAST and uses rpsblast results against COG (Clusters of Orthologous Groups) to leverage alignments and avoid exhaustive all-versus-all BLAST.
• Sequence clustering: Performs sequence clustering to reduce the number of comparisons and overall computational demands.
• Performance: Reduces computational time while maintaining high accuracy in identifying protein family members.

Scientific Applications:

• Protein family identification: Identification and confirmation of protein family members across large sequence databases using PSI-BLAST followed by HMMer validation.
• Functional annotation: Annotating proteins via rpsblast matches to COG (Clusters of Orthologous Groups) and alignments from FastBLAST.
• Large-scale comparative analysis: Enabling large-scale comparative and clustering analyses that would otherwise require all-versus-all BLAST or InterProScan.

Methodology:

Uses PSI-BLAST for initial candidate detection, confirms hits with HMMer profile HMM searches, and can leverage FastBLAST alignments plus rpsblast results against COG and sequence clustering to avoid all-versus-all BLAST.