HMM-FRAME

HMM-FRAME corrects sequencing-induced frameshifts and improves protein domain classification in metagenomic datasets by modeling insertions and deletions in homopolymer regions typical of pyrosequencing reads.

Key Features:

• Augmented Viterbi Algorithm: Employs an augmented Viterbi algorithm that incorporates sequencing-platform-specific error models for alignment to profile HMMs.
• Error Correction and Classification: Detects and corrects frameshifts caused by insertions and deletions and classifies putative gene fragments into their respective domain families with increased sensitivity and specificity.
• Enhanced Alignment Quality: Produces longer alignments with smaller E-values compared to conventional profile HMM alignment methods after frameshift correction.
• Application in Metagenomics: Applied to targeted and published metagenomic datasets to improve domain classification in studies with prevalent sequencing-induced frameshifts.

Scientific Applications:

• Complement to profile HMM methods: Rescues marginal profile HMM alignments affected by frameshifts to improve detection of protein domains.
• Metagenomic protein annotation: Enhances accuracy of protein domain classification and functional annotation in metagenomic sequencing projects.
• Pyrosequencing error handling: Recovers coding-frame information in reads with homopolymer-associated indels typical of pyrosequencing.

Methodology:

Integrates sequencing-platform-specific error models into an augmented Viterbi algorithm to model insertions and deletions during alignment to profile HMMs, enabling frameshift correction that yields longer alignments and reduced E-values.