To create a nucleotide sequence from a protein document, go to Sequence → Back Translate. Ambiguous back-translation uses a specific genetic code to produce a nucleotide sequence with ambiguous bases, so that every possible codon is represented for each amino acid. Un-ambiguous back-translation uses codon usage tables to produce a nucleotide sequence where the most frequently used codon for that organism is used for each amino acid. Codon usage tables for some organisms are provided, or you can you can import custom codon usage tables in GCG CodonFrequency and EMBOSS cusp formats. For further instructions, see How do I create a custom codon usage table?.