Closed cdanielmachado closed 6 years ago
Yup, Moritz and I discussed this already. We've implemented the method from [1], which I extended by the menaquinols I found on BiGG. You make a good point that we could look for universal cofactors instead, we could also make these run only if the cofactors can be identified in the model and skip those that we cannot identify. I think we should do both :D Extend our range of tests and skip if the cofactors are unavailable.
[1] Fritzemeier, C. J., Hartleb, D., Szappanos, B., Papp, B., & Lercher, M. J. (2017). Erroneous energy-generating cycles in published genome scale metabolic networks: Identification and removal. PLoS Computational Biology, 13(4), 1–14. http://doi.org/10.1371/journal.pcbi.1005494
I would just test the balance of the cofactors that are present in the model, and do not make any assumptions of which ones "should" be there.
This is now resolved by using MetaNetX IDs and skipping co-factors not in the model.
I am getting errors for testing production "mql6" and "mql7" because they are not in the model.
According to the paper of Xavier et al [1], these cofactors are not universal. Therefore, this should not be an error. If they are not in the model then maybe the model is just fine.
[1] http://www.sciencedirect.com/science/article/pii/S1096717616302415