To effectively predict mucins from genomic and protein sequences obtained from genome projects, we developed a strategy based on the amino acid compositional bias characteristic of the mucin domains.