Development of a Knowledge Graph-Driven Machine Learning Model for Personalised Treatment Recommendations in Cystic Fibrosis (CF)

A new machine learning model will use a knowledge graph—a structured map of biological relationships—to recommend personalised treatments for people with cystic fibrosis (CF). CF is a complex genetic condition caused by mutations in the CFTR gene, which primarily affects the lungs and digestive system. Treatment responses vary widely between individuals, yet current approaches often rely on trial-and-error prescribing. This project addresses that gap by building a model that integrates diverse data—genetic mutations, clinical outcomes, drug interactions—into a single predictive framework. The goal is to move from one-size-fits-all guidelines to recommendations tailored to each patient’s specific biology. If successful, this could change how clinicians decide which therapies to try first, potentially reducing the time patients spend on ineffective treatments. The impact would be felt in everyday clinical practice: faster access to the right drug, fewer side effects, and better long-term management of a progressive disease. This is applied research with a clear practical endpoint—not fundamental science—so the benefits, if realised, would be direct and measurable in hospital settings.