Bacterial modular polyketide synthases are multi-functional enzymes that mediate the assembly of a wide variety of complex natural products. Many of these metabolites have been developed into clinically approved antibiotics, anti-parasitic agents, and cancer therapeutics. For the past three decades numerous groups have worked to understand the function and structural parameters involved in expanding the structural diversity of modular PKSs through in vivo and in vitro approaches. Although some successes have been reported, there continues to be a need to expand our knowledge about the role of individual catalytic domains, protein-protein interactions, and substrate selectivity parameters in order to engineer these systems. We have focused on studies of 12-, 14- and 16-membered macrolactone generating PKSs to understand the primary bottlenecks in production of new metabolites using unnatural chain elongation intermediates. A particular focus in these studies has been the PKS terminal thioesterase domain and the determinants that contribute to selectivity and macrolactonization. Recent studies have revealed the ability to maximize efficiency of macrolactone formation by appropriate pairing of a TE domain in an engineered PKS involved in processing of unnatural substrates. We have also significantly expanded our pool of unnatural substrates to assess flexibility of PKSs toward proximal/distal functional group alterations, heteroatom replacements, and odd-numbered chain elongation intermediates. In numerous examples, new 12-, 14- and 16-membered ring macrocycles are generated and can be further transformed into novel macrolides by glycosylation and late-stage oxidation with heterologous CYP450s. This biocatalytic cascade approach is enabling scalable methods for understanding PKS function, engineering biosynthetic enzymes and obtaining new bioactive molecules for analysis against a range of disease targets.