%0 Journal Article %1 Hartleb2016Improved %A Hartleb, Daniel %A Jarre, Florian %A Lercher, Martin J. %D 2016 %I Public Library of Science %J PLOS Computational Biology %K flux-analysis gaps gene-deletion %N 8 %P e1005036+ %R 10.1371/journal.pcbi.1005036 %T Improved Metabolic Models for E. coli and Mycoplasma genitalium from GlobalFit, an Algorithm That Simultaneously Matches Growth and Non-Growth Data Sets %U http://dx.doi.org/10.1371/journal.pcbi.1005036 %V 12 %X Constraint-based metabolic modeling methods such as Flux Balance Analysis (FBA) are routinely used to predict the effects of genetic changes and to design strains with desired metabolic properties. The major bottleneck in modeling genome-scale metabolic systems is the establishment and manual curation of reliable stoichiometric models. Initial reconstructions are typically refined through comparisons to experimental growth data from gene knockouts or nutrient environments. Existing methods iteratively correct one erroneous model prediction at a time, resulting in accumulating network changes that are often not globally optimal. We present GlobalFit, a bi-level optimization method that finds a globally optimal network, by identifying the minimal set of network changes needed to correctly predict all experimentally observed growth and non-growth cases simultaneously. When applied to the genome-scale metabolic model of Mycoplasma genitalium, GlobalFit decreases unexplained gene knockout phenotypes by 79\%, increasing accuracy from 87.3\% (according to the current state-of-the-art) to 97.3\%. While currently available computers do not allow a global optimization of the much larger metabolic network of E. coli, the main strengths of GlobalFit are already played out when considering only one growth and one non-growth case simultaneously. Application of a corresponding strategy halves the number of unexplained cases for the already highly curated E. coli model, increasing accuracy from 90.8\% to 95.4\%. Mathematical models that aim to describe the complete metabolism of a cell help us understand cellular metabolic capabilities and evolution, and aid the biotechnological design of microbial strains with desired properties. Draft models are frequently improved through adjustments that increase the agreement of growth/non-growth predictions with observations from gene knockout experiments. Automated methods for this task typically correct one erroneous prediction after the other. We present GlobalFit, a novel method that can consider all experiments and all possible changes simultaneously to identify model modifications that are globally optimal (i.e., that correct the largest possible number of wrong predictions while introducing sets of changes that are most compatible with existing knowledge). This becomes computationally very hard when considering large metabolic models; however, a reduced application of GlobalFit that only looks at small subsets of experiments simultaneously works very well in practice. Allowing only changes that are conservative (e.g., introducing new reactions only if supported by significant genomic evidence), GlobalFit halves the number of wrong growth/non-growth predictions for the state-of-the-art metabolic models of E. coli and Mycoplasma genitalium, increasing prediction accuracy to 95.4\% and 93.0\%, respectively. By additionally allowing less conservative changes, we are able to improve accuracy further to 97.3\% for the M. genitalium model.

@article{Hartleb2016Improved, abstract = {Constraint-based metabolic modeling methods such as Flux Balance Analysis ({FBA}) are routinely used to predict the effects of genetic changes and to design strains with desired metabolic properties. The major bottleneck in modeling genome-scale metabolic systems is the establishment and manual curation of reliable stoichiometric models. Initial reconstructions are typically refined through comparisons to experimental growth data from gene knockouts or nutrient environments. Existing methods iteratively correct one erroneous model prediction at a time, resulting in accumulating network changes that are often not globally optimal. We present {GlobalFit}, a bi-level optimization method that finds a globally optimal network, by identifying the minimal set of network changes needed to correctly predict all experimentally observed growth and non-growth cases simultaneously. When applied to the genome-scale metabolic model of Mycoplasma genitalium, {GlobalFit} decreases unexplained gene knockout phenotypes by 79\%, increasing accuracy from 87.3\% (according to the current state-of-the-art) to 97.3\%. While currently available computers do not allow a global optimization of the much larger metabolic network of E. coli, the main strengths of {GlobalFit} are already played out when considering only one growth and one non-growth case simultaneously. Application of a corresponding strategy halves the number of unexplained cases for the already highly curated E. coli model, increasing accuracy from 90.8\% to 95.4\%. Mathematical models that aim to describe the complete metabolism of a cell help us understand cellular metabolic capabilities and evolution, and aid the biotechnological design of microbial strains with desired properties. Draft models are frequently improved through adjustments that increase the agreement of growth/non-growth predictions with observations from gene knockout experiments. Automated methods for this task typically correct one erroneous prediction after the other. We present {GlobalFit}, a novel method that can consider all experiments and all possible changes simultaneously to identify model modifications that are globally optimal (i.e., that correct the largest possible number of wrong predictions while introducing sets of changes that are most compatible with existing knowledge). This becomes computationally very hard when considering large metabolic models; however, a reduced application of {GlobalFit} that only looks at small subsets of experiments simultaneously works very well in practice. Allowing only changes that are conservative (e.g., introducing new reactions only if supported by significant genomic evidence), {GlobalFit} halves the number of wrong growth/non-growth predictions for the state-of-the-art metabolic models of E. coli and Mycoplasma genitalium, increasing prediction accuracy to 95.4\% and 93.0\%, respectively. By additionally allowing less conservative changes, we are able to improve accuracy further to 97.3\% for the M. genitalium model.}, added-at = {2018-12-02T16:09:07.000+0100}, author = {Hartleb, Daniel and Jarre, Florian and Lercher, Martin J.}, biburl = {https://www.bibsonomy.org/bibtex/2cb91cdc68b2e836c497318d7cc06a6c4/karthikraman}, citeulike-article-id = {14380723}, citeulike-linkout-0 = {http://dx.doi.org/10.1371/journal.pcbi.1005036}, day = 2, doi = {10.1371/journal.pcbi.1005036}, interhash = {d7797b56327585ba95ca65b9ab2962a4}, intrahash = {cb91cdc68b2e836c497318d7cc06a6c4}, journal = {PLOS Computational Biology}, keywords = {flux-analysis gaps gene-deletion}, month = aug, number = 8, pages = {e1005036+}, posted-at = {2017-06-22 06:37:11}, priority = {2}, publisher = {Public Library of Science}, timestamp = {2018-12-02T16:09:07.000+0100}, title = {Improved Metabolic Models for E. coli and Mycoplasma genitalium from {GlobalFit}, an Algorithm That Simultaneously Matches Growth and {Non-Growth} Data Sets}, url = {http://dx.doi.org/10.1371/journal.pcbi.1005036}, volume = 12, year = 2016 }