Christos Mitsanis

From Diverse Prior Knowledge to Mechanistic Causal Network Using PSoup: A Case Study in Shoot Branching

The paper develops a parameter-free model of shoot branching using a causal network derived from literature, achieving 86% accuracy in predicting branching direction from 78 perturbations.

FLASH-P: Turning decades of biology into accurate causal networks with AI agents

FLASH-P constructs causal regulatory networks from literature, predicting perturbation outcomes with 90% accuracy across seven species.

PSoup: an R package for simulating biological networks from a qualitative perspective

PSoup is an R package that converts knowledge graphs into qualitative, parameter-free models to predict biological network behavior under perturbations.