Abstract: Methods are disclosed for the design of non-native (i.e. heterologous) polypeptides comprising a proportion of hydrophobic amino acids which have an increased probability of being efficiently expressed in an expression system such as a bacterial host (e.g. E. coli). The methods involve identifying one or more hydrophobic peptide sequences within a polypeptide of interest, and arranging or re-locating at least one of the hydrophobic peptide sequences within said polypeptide so as to generate a candidate polypeptide with reduced amplitude in hydrophobicity and/or length of any hydrophobic region(s). Such methods are particularly useful for designing polyepitope polypeptides, and specific examples of such are described for Epstein-Barr virus (EBV), hepatitis C virus (HCV) and human immunodeficiency virus (HIV).