Optimal synthesis of protein purification processes

There has been an increasing interest in the development of systematic meth ods for the synthesis of purification steps for biotechnological products, which are often the most difficult and costly stages in a biochemical proce ss. Chromatographic processes are extensively used in the purification of m ulticomponent biotechnological systems. One of the main challenges in the s ynthesis of purification processes is the appropriate selection and sequenc ing of chromatographic steps that are capable of producing the desired prod uct at an acceptable cost and quality. This paper describes mathematical mo dels and solution strategies based on mixed integer linear programming (MIL P) for the synthesis of multistep purification processes. First, an optimiz ation model is proposed that uses physicochemical data on a protein mixture , which contains the desired product, to select a sequence of operations wi th the minimum number of steps from a set of candidate chromatographic tech niques that must achieve a specified purity level. Since several sequences that have the minimum number of steps may satisfy the purity level, it is p ossible to obtain the one that maximizes final purity. Then, a second model that may use the total number of steps obtained in the first model generat es a solution with the maximum purity of the product. Whenever the sequence does not affect the final purity or more generally does not impact the obj ective function, alternative models that are of smaller size are developed for the optimal selection of steps. The models are tested in several exampl es, containing up to 13 contaminants and a set of 22 candidate high-resolut ion steps, generating sequences of six operations, and are compared to the current synthesis approaches.