PfamA

PfamA is the manually curated component of the Pfam protein families database. Pfam catalogs protein families, domains, and motifs by providing high-quality multiple sequence alignments and profile hidden Markov models (HMMs) that describe conserved patterns within families. Each Pfam-A entry represents a curated family and is identified by an accession code such as PFxxxxx.

An Pfam-A entry includes a seed alignment, a full alignment, a curated description of the family’s function, and information about the domain architecture of member proteins. From the seed alignment, a profile HMM is built using the HMMER software, which is then used to detect and annotate homologous sequences. The full alignment expands representation of the family beyond the seed to include additional related sequences. Pfam-A entries carry cross-references to other resources and often include notes on structure, motifs, and conserved residues.

Pfam-A models are widely used to annotate protein sequences, identify conserved domains, and determine domain architectures across genomes. They support functional inference and comparative analyses, and are integrated with other resources such as InterPro, UniProt, and PDB. The Pfam database is updated regularly to incorporate new data and refinements of already curated families. Pfam has historically included Pfam-B, an automatically generated set of additional families that complemented Pfam-A, though Pfam-A remains the core of curated protein family coverage.