Distinguishing Context Dependent Events in Quotients of Causal Stories
Résumé
Causality analysis of rule-based models allows the reconstruction of the causalpaths leading to chosen events of interest. This potentially reveals emerging paths that werecompletely unknown at the time of creation of a model. However, current implementationsprovide results in the form of a collection of stories. For large models, this can amount tohundreds of story graphs to read and interpret for a single event of interest. In this work,we hence develop a method to fold a collection of stories into a single quotient graph.The main challenge is to find a trade-off in the partitioning of story events which willmaximize compactness without loosing important details about information propagation inthe model. The partitioning criterion proposed is relevant context, the context from anevent’s past which remains useful in its future. Each step of the method is illustrated on atoy rule-based model. This work is part of a longer term objective to automatically extractbiological pathways from rule-based models.
Origine | Fichiers produits par l'(les) auteur(s) |
---|