Release Notes for PyCoMo 0.2.0
New Features
Loopless FVA added to PyCoMo community metabolic models.
Thermodynamically infeasible cycles (or loops) lead to inflated flux values for reactions that are part of them. Such loops are expected to occur when forming a community metabolic model, when community members share a set of external metabolites that can be reversibly converted. In this case, all directions of the reactions are in principle allowed, but the reactions running in loops have to be avoided. Solutions have been proposed and implemented for this problem, e.g. in cobrapy 1, which uses the concepts of CycleFreeFlux 2.
However, the implementation of COBRApy cannot be used with the model structure of PyCoMo. Therefore, loopless FVA based on CycleFreeFlux, tailored to PyCoMo has been implemented.
A tutorial can be found in the tutorial directory.
Parallel implementation of FVA and loopless FVA
FVA of compartmentalized community metabolic models requires a lot of time, especially for genome-scale models with more than a handfull members. To make better use of available resources, FVA and loopless FVA are now set up to use multiple processes. Parallel processing is enabled by using a COBRApy configuration object, or by specifying the number of processes in the FVA function call.
Computing the maximum growth rate across all community compositions
A function for computing the overall maximum growth-rate of a community metabolic model has been added. A tutorial can be found in the tutorial directory.
Fixes
- When constructing community metabolic models, any previous constraints are lifted on reactions transporting metabolites between the shared medium compartment and the members. This prevents media constraints for members to spill over to the community metabolic model.
-
CycleFreeFLux: Desouki et al., CycleFreeFlux: efficient removal of thermodynamically infeasible loops from flux distributions, Bioinformatics, Volume 31, Issue 13, July 2015, Pages 2159–2165, https://doi.org/10.1093/bioinformatics/btv096 ↩