Addressing Network Medicine Challenges: an example of causal examination of potential mechanisms of metformin’s impact on C9orf72-mediated ALS using tools of advanced knowledge management platform

Data integration from structured, semi-structured or unstructured sources often of different origin, different granularity and different quality is one of main challenges of wider use of network medicine frameworks. But the biggest challenge seems to lie in complexity of biological systems, their heterogeneity, including interindividual variability, and particularly in network inferences, causal relationships and their dynamic character.

Valuable tools addressing these important challenges are causal graphs modelling complex relationships, identifying key causal factors, enabling applications of causal inference methods, contrafactual reasoning, hypotheses generation and easier interpretable representation of complex biological systems and their systematic analysis. In the field of personalized medicine causal graphs enable understanding the causal relationships of important complex processes, predictions about disease progression, assessment individual or population-level risks, helping design targeted interventions and their monitoring.

Advanced knowledge platforms might be of a big importance in the field of drug repurposing by addressing network medicine challenges. In our work we used advanced knowledge platform bigRing™ providing no-code functionalities of universal semantic and metaheuristic model explicitly expressing the knowledge in unique model of reality interpretation containing dynamically and causally (cause and effect) interconnected network-elements into multidimensional dynamic structure ⁴.

We tackled described challenges by using “small-scale” biological systemic data acquired from KEGG to the platform and analyzed by bigRing’s no-code model designing tool, graphic database and AI engine. We used relevant datasets as well as proprietary data. The multidimensional dynamically structured data were examined with focus on metabolic pathways through which metformin might influence C9orf72-mediated ALS, the most prevalent genetic form ¹. We were inspired in part by insights suggesting that metformin can modulate PKR activation and subsequently reduce RAN protein levels in C9orf72 models. Use of these tools provided meaningful insights based on causal inferences. Some other pathways being reported as participating on ALS pathogenesis as mTOR, AMPK were examined.

Our findings suggest that metformin holds promise as a potential treatment for patients with this C9orf72 medited ALS variant. Furthermore, bigRing emerges as a valuable tool for identifying novel therapeutic approaches, especially in the realm of drug-repurposing.