TIGRFAMs is a manually curated database of protein families, represented by hidden Markov models (HMMs) and associated information such as multiple sequence alignments, Gene Ontology (GO) assignments, and literature references. The database aims to support both automated and manual genome annotation. TIGRFAMs includes models of full-length proteins and shorter regions, categorized into superfamilies, subfamilies, and equivalogs. Equivalogs are sets of homologous proteins conserved in function since their last common ancestor. The scope of each model is adjusted by setting cutoff scores and selecting seed alignment members based on sequence similarity, functional evidence, phylogenetics, metabolic context, and neighboring genes. TIGRFAMs complements Pfam, which focuses on broad coverage of distant homologs. The current release contains over 1600 protein families, with 837 classified as equivalogs. TIGRFAMs is available for searching and downloading at www.tigr.org/TIGRFAMs.TIGRFAMs is a manually curated database of protein families, represented by hidden Markov models (HMMs) and associated information such as multiple sequence alignments, Gene Ontology (GO) assignments, and literature references. The database aims to support both automated and manual genome annotation. TIGRFAMs includes models of full-length proteins and shorter regions, categorized into superfamilies, subfamilies, and equivalogs. Equivalogs are sets of homologous proteins conserved in function since their last common ancestor. The scope of each model is adjusted by setting cutoff scores and selecting seed alignment members based on sequence similarity, functional evidence, phylogenetics, metabolic context, and neighboring genes. TIGRFAMs complements Pfam, which focuses on broad coverage of distant homologs. The current release contains over 1600 protein families, with 837 classified as equivalogs. TIGRFAMs is available for searching and downloading at www.tigr.org/TIGRFAMs.