Polyketide synthases (PKSs) are huge multi-enzymatic protein complexes involved in the biosynthesis of one of the largest families of bioactive natural products, namely polyketides. The specificity of interactions between various catalytic domains of these megasynthases is one of the pivotal factors which control the precise order in which the extender units are joined during the biosynthetic process. Hence, understanding the molecular details of protein-protein interactions in the PKS megasynthases would be crucial for rational design of novel polyketides by domain swapping experiments involving engineered combinations of PKS catalytic domains. We have developed a computational method for exploring the binding interface between two proteins, and used it to identify the interacting residue pairs, which govern the specificity of recognition between acyl carrier protein (ACP) domain and two core catalytic domains, namely the ketosynthase (KS) and acyl transferase (AT). Both of these domain interactions i.e. the KS-ACP and the AT-ACP, are likely to play a major role in channelling of substrates and control of specificity during polyketide biosynthesis. The method, called interface scan, uses a combination of geometric docking and evolutionary information for the identification of the most appropriate mode of association between two proteins. The parameters of interface scan have been standardized based on analysis of contacts in the crystal structure of ACP in complex with ACP synthase (AcpS). Many of the contacts predicted for PKS domains are in agreement with available experiments.