Attempts at generating novel chemistries by genetically manipulating polyketide synthases (PKSs) usually result in no detectable or poor product yield. Understanding processes that drive the evolution of PKSs might provide a solution to this problem. The synonymous-to-non-synonymous nucleotide substitution ratios across alignments of well-characterized PKS modules were examined using a sliding windows approach. Not surprisingly, the overall substitution ratios showed that PKS modules are generally under strong purifying selection, confirming experimental observations that changes to the primary amino acid sequence, regardless of whether these changes are conservative or not, will most likely result in some loss in function. Despite the masking effect of negative selection, by judicious choice of window size, it was possible to recognize amino acid residues that appear to be under strong positive selection. The importance of these amino acids has not been recognized by other analysis methods before and we suggest that they may function to 'fine tune' modular PKSs. Future efforts will concentrate on understanding if this 'fine tuning' is at the level of protein expression, for example, transcription or translation, or at the level of protein function, for example, efficient selection and channeling of acyl intermediates between domains.