TY - JOUR TI - Phenotype-based cell-specific metabolic modeling reveals metabolic liabilities of cancer AU - Yizhak, Keren AU - Gaude, Edoardo AU - Le Dévédec, Sylvia AU - Waldman, Yedael Y AU - Stein, Gideon Y AU - van de Water, Bob AU - Frezza, Christian AU - Ruppin, Eytan A2 - Dang, Chi Van VL - 3 PY - 2014 DA - 2014/11/21 SP - e03641 C1 - eLife 2014;3:e03641 DO - 10.7554/eLife.03641 UR - https://doi.org/10.7554/eLife.03641 AB - Utilizing molecular data to derive functional physiological models tailored for specific cancer cells can facilitate the use of individually tailored therapies. To this end we present an approach termed PRIME for generating cell-specific genome-scale metabolic models (GSMMs) based on molecular and phenotypic data. We build >280 models of normal and cancer cell-lines that successfully predict metabolic phenotypes in an individual manner. We utilize this set of cell-specific models to predict drug targets that selectively inhibit cancerous but not normal cell proliferation. The top predicted target, MLYCD, is experimentally validated and the metabolic effects of MLYCD depletion investigated. Furthermore, we tested cell-specific predicted responses to the inhibition of metabolic enzymes, and successfully inferred the prognosis of cancer patients based on their PRIME-derived individual GSMMs. These results lay a computational basis and a counterpart experimental proof of concept for future personalized metabolic modeling applications, enhancing the search for novel selective anticancer therapies. KW - genome-scale metabolic modeling KW - cancer KW - selective drug target KW - personalized medicine JF - eLife SN - 2050-084X PB - eLife Sciences Publications, Ltd ER -