Modeling rare metabolic diseases with reaction systems

Inherited rare disorders of phenylalanine–tyrosine (Phe–Tyr) catabolism are difficult to analyze computationally because overlapping clinical phenotypes often mask the underlying genetic cause, while comparable kinetic data are unavailable for these rare diseases. To address this, we have developed a qualitative model based on Reaction Systems (RSs) and Discrete Entity level Annotation (DEA). The DEA method encodes enzymes into discrete levels describing their functioning (normal, hypomorphic, null), and metabolites and cofactors into qualitative levels (low, normal, accumulated). The resulting model comprises a set of manually curated rules that capture seven physiological transformations and thirty-two pathological mechanisms, describing enzyme dysfunction, metabolite accumulation, reduced downstream production, and alternative routes. We have implemented our model in a computational tool for RSs, called BioReSolve, that we have previously designed. Starting from the computation graph we have studied the attractors of the system and have performed causal analyses for the derivation of the target molecules. We have considered different target molecules corresponding to biomarkers associated with specific metabolic disorders. Our methodology provides a formal representation of phenylalanine–tyrosine pathway. By our analysis, we can capture multi-enzyme alterations and may also support diagnostic reasoning when detailed quantitative information is unavailable.