Simulate distribution of 13C for a given metabolic network at dynamic equilibrium (= constant metabolite pool sizes).
MetSim is employing object oriented programming where each metabolite pool is an object. Metabolite pools are stored in pandas dataframes where each column represents a carbon and each row a molecule. Unlabeled (12C) carbons are represented by a 0, while labeled (13C) carbons are represented by a 1. Consequently, a molecule is a sequence of 0s and 1s, for example {'C1': 0, 'C2': 0, 'C3': 1} represents a three-carbon metabolite which is labeled once at the C3 position.
Calculation is done step-wise and every step consists of two sub-steps. First, from every metabolite pool random molecules are moved to an intermediate pool called tmp and subsequently distributed to the specified neighboring metabolite pools.
All necessary sub-steps are defined as class methods or functions and ultimately strung together to form a sequence which makes up a step which is iterated to simulate the metabolism. Enrichment is calculated, stored in a dataframe and can afterwards be saved to csv-files and plotted.
The following metabolic network depicting the central energy metabolism as well as selected amino acids was used as an example.
10 000 steps were simulated in roughly 20 minutes. As an example, the enrichment and isotopologue distribution of aspartate is shown, all other results can be viewed in the examples folder.
