Amino acid residue misincorporations are necessarily found in sequence variants at low concentrations in admixture with expressed polypeptides, resulting from one or more base mismatches within codons susceptible to amino acid residue misincorporation during transcription and/or translation. The invention provides a method of optimizing the coding sequences of a polynucleotide that encodes a polypeptide, wherein at least one codon is susceptible to amino acid residue misincorporation. The method of the invention can be used to reverse-engineer an unknown coding sequence, which encodes the same polypeptide, but differs in said at least one codon from the known coding sequence. The method can further be used to alter the immunogenic potential of an expressed polypeptide. Thus, the invention is useful in engineering optimized polynucleotides encoding polypeptides.