An algorithm for removing stoichiometric discrepancies in biochemical reaction databases

摘要

It is observed that a significant number of reactions are stoichiometrically unbalanced in the existing databases (e.g. MetaCyc), even in curated databases like RiceCyc. To obtain a valid model, reactions should be mass balanced. Here, we propose a systematic algorithm based on generalized inverse matrix - to search all the reactions where mass is not conserved and to calculate the appropriate balancing coefficients of reactants and products, wherever possible (feasible reactions). Using this algorithm we have identified the set of reactions which cannot be stoichiometrically balanced with the present set of substrates and products (infeasible reactions). We have also suggested names of filler metabolites to turn infeasible reactions into feasible reactions.