TIGRFAMS protein family database

TIGRFAMS protein family database provides curated protein family definitions using hidden Markov models (HMMs) built from seed alignments to enable automated annotation and functional prediction across bacterial and archaeal genomes.

Key Features:

• Hidden Markov Models (HMMs): Protein families are represented by HMMs constructed from curated seed alignments of trusted representative sequences.
• Seed alignments: Each TIGRFAMs entry includes a seed alignment of trusted representative sequences that forms the basis for the corresponding HMM.
• Cutoff scores: HMM-derived cutoff scores are used to determine membership of proteins in specific families.
• Annotation transfer: Annotations can be transferred to member proteins identified by HMM matches to support functional prediction.
• Model types: Entries are categorized as equivalog, subfamily, or domain models to reflect conserved function, subfamily-level variation, or domain-level annotations.
• Integration with Genome Properties: HMM results from TIGRFAMs are used as computed evidence in the Genome Properties database for subsystem assessment.
• Focus on bacteria and archaea: The resource primarily targets bacterial and archaeal subsystems.
• Content scale: TIGRFAMs contains 4284 models and Genome Properties describes 628 systems.
• Curatorial sources: Content is derived from subsystem discovery efforts and biocuration of the scientific literature.
• Use in automated pipelines: TIGRFAMs HMMs are applied in automated annotation pipelines to assign family membership and propagate annotations.

Scientific Applications:

• Protein family assignment: Assigns proteins to families for functional annotation using HMM matches and cutoff scores.
• Functional prediction: Enables transfer of functional annotations to member proteins identified by HMMs.
• Subsystem reconstruction: Provides computed evidence for reconstructing molecular subsystems such as enzymatic pathways and protein complexes via Genome Properties.
• Comparative genomics: Supports analysis of presence/absence and distribution of protein families and subsystems across bacterial and archaeal genomes.

Methodology:

Seed alignments of trusted representative sequences are used to build HMMs; HMM matches with defined cutoff scores determine family membership and enable annotation transfer; Genome Properties uses TIGRFAMs HMM results as computed evidence to assess molecular subsystems, and subsystem reconstruction across genomes informs selection of seed sequences and cutoff values.