The Maranas group is working on the development of algorithmic and, in particular, optimization techniques to support the analysis and redesign of biological systems at different scales. At the protein level, we are interested in computationally inferring what amino acid compositions are likely to yield i) proteins or antibodies with targeted binding affinities and ii) enzymes with improved stability, specificity and activity for specific biotransformations. To this end, we make use of ab initio energy calculations at the ground and transition states, MD simulations, as well as scoring functions based on bioinformatics inspired analyses. At the metabolic network level, we are pursuing methods for automating the generation, curation, and correction of genome-scale models of metabolism. We are also interested in generating isotope mapping models to support metabolic flux elucidation using MFA. In addition, we are working towards developing computational tools to help decide how to engineer (i.e., through gene knock-in/out/up/down(s)) biological production systems. A unifying feature of these seemingly disjoint research targets is the need to systematically search through many network configurations, amino acid compositions, protein structures, etc. and identify the "best" one. To this end, the development of efficient theoretical, algorithmic, and computational techniques for arriving at relevant as well as theoretically sound results while maximizing computational efficiency is pursued.