Biological pathways have been extensively analyzed and modeled in respect to overall conversions of substrates to products, although information on atom transitions or atom flow in metabolic networks is not available in databases. The reaction atom mapping designates the track of all the atoms from substrates to products in each individual reaction and consequently, it elucidates the mass flow in a pathway of reactions. The study of metabolism at the atomic level is of great importance in many applications of systems biology such as simulation of isotope labeling experiments and pathway inference in metabolic engineering. The automated identification of atom transition within a pathway of reactions is a challenging task since the degree of complexity of metabolic networks dramatically increases when we pass from “metabolite-level studies” to “atom-level studies”. Despite being extensively studied in various approaches, the field of atom mapping of metabolic networks is lacking an automated approach, which takes into account the information of reaction mechanism for atom mapping and is easily extendable from “atom-mapped reaction” to “atom-mapped pathway of reactions”. Hereby we introduce a computational framework for the reconstruction of metabolic networks from in silico labeled substrates, which results in atom-level representation of metabolic networks and allows the atom tracking through the resulted networks. Our method allows a straightforward and computationally efficient means for the observation of all the possible fluxes in a proposed network model based on atom biotransformation. We illustrate the efficacy and the application of our method through the reconstruction of E-coli atom-mapped reaction network.