The gene set for marmoset was built using the Ensembl gene annotation pipeline. Species-specific resources for marmoset are relatively limited and therefore sequences from other primates, especially human, were also used as supporting evidence for coding transcript models. In addition to the coding transcript models, non-coding RNAs and pseudogenes were annotated. Marmoset cDNAs and ESTs and human cDNAs were mapped to the genome using Exonerate.