Computational evaluation of cellular metabolic costs successfully predicts genes whose expression is deleterious

• DOI: 10.1073/pnas.1312361110
• Publication Date: 2013-11-07T04:11:21Z
• Authors: Allon Wagner; Raphy Zarecki; Leah Reshef; Camelia Gochev; Rotem Sorek; Uri Gophna; Eytan Ruppin
• Keywords: 570; Neoplastic; Models, Genetic; Gene Expression Profiling; 0206 medical engineering; flux balance analysis; Computational Biology; Gene Expression; 02 engineering and technology; systems metabolic engineering; 3. Good health; Gene Expression Regulation, Neoplastic; metabolic modeling; Gene Expression Regulation; Genetic; constraint-based modeling; Models; horizontal gene transfer; Humans; Algorithms; Metabolic Networks and Pathways

Significance Biologists frequently overexpress genes to learn about their cellular functions, and biotechnologists do so to construct novel metabolic pathways that produce valuable chemical compounds. However, gene overexpression often leads to deleterious consequences whose cause is unclear. Here, we present a computational method named Expression-Dependent Gene Effects (EDGE) that can successfully predict the deleterious effects resulting from overexpression of either native or foreign (originating in another species) metabolic genes. EDGE relies on genome-scale metabolic models, an emerging computational paradigm for studying metabolism in silico. Beyond its biotechnological significance, gene overexpression also plays an important role in human disease. We show EDGE’s applicability in the latter case by demonstrating its ability to detect toxic genes whose expression tends to be suppressed in cancer cells.