| Popis: |
MotivationThe growing complexity of reaction-based models necessitates early detection and resolution of model errors. This paper addresses mass balance errors, discrepancies between the mass of reactants and products in reaction specifications. One approach to detection is atomic mass analysis, which uses meta-data to expose atomic formulas of chemical species. Atomic mass analysis isolates errors to individual reactions. However, this approach burdens modelers with expressing model meta-data and writing excessively detailed reactions that include implicit chemical species (e.g., water). Moreover, atomic mass analysis has the shortcoming of not being applicable to large molecules because of the limitations of current annotation techniques. The second approach, Linear Programming (LP) analysis, avoids using model meta-data by checking for a weaker condition, stoichiometric inconsistency. But this approach suffers from false negatives and has large isolation sets (the set of reactions implicated in the stoichiometric inconsistency).ResultsWe propose alternatives to both approaches. Our alternative to atomic mass analysis is moiety analysis. Moiety analysis uses model meta-data in the form of moieties present in chemical species. Moiety analysis avoids excessively detailed reaction specifications, and can be used with large molecules. Our alternative to LP analysis is Graphical Analysis of Mass Equivalence Sets (GAMES). GAMES has a slightly higher false negative rate than LP analysis, but it provides much better error isolation. In our studies of the BioModels Repository, the average size of isolation sets for LP analysis is 55.5; for GAMES, it is 5.4. We have created open source codes for moiety analysis and GAMES.Availability and ImplementationOur project is hosted at https://github.com/ModelEngineering/ SBMLLint, which contains examples, documentation, source code files, and build scripts used to create SBMLLint. Our source code is licensed under the MIT open source license.Contact:jlheller@uw.eduSupplementary information: None. |
