HMMER

HMMER is a bioinformatics software package for sequence analysis that uses profile hidden Markov models (HMMs) to detect homologous sequences and conserved domains. It was developed by Sean R. Eddy and is widely used in protein family annotation and genome analysis. The current major release, HMMER3, emphasizes speed and sensitivity for large-scale data sets and is distributed as open-source software.

Core tools in the HMMER suite include hmmbuild, which constructs a profile HMM from a multiple sequence alignment; hmmalign, which aligns sequences to a profile HMM; hmmpress, which creates an indexed database for fast searches; hmmsearch, which searches a sequence database with a profile HMM; and hmmcalibrate to estimate statistical parameters. Additional programs such as jackhmmer perform iterative searches, nhmmer searches DNA with DNA-profile HMMs, and phmmer and hmmscan provide other search modes.

HMMER reports sequence similarity using log-odds scores, E-values, and bit scores; search results include alignments and domain architectures. It relies on probabilistic algorithms, including the Viterbi path and Forward–Backward computations, with acceleration through prefiltering and optimized code to handle large databases. HMMER models are typically built from curated multiple sequence alignments and are used by major protein family resources such as Pfam for annotation.

History and availability: HMMER originated in the mid-1990s, with HMMER3 released in the early 2010s as a major rewrite; it remains actively developed. It is cross-platform (Linux, macOS, Windows via ports) and released under an open-source license. The software is widely distributed through the HMMER project website and repositories and is a staple in many bioinformatics pipelines.