Graphs for Causal Discovery (Graphs4Causal)
Project ID: 2531bd1706
(You will need this ID for your application)
Research Theme: Mathematical Sciences
Research Area(s): mathematical sciences
UCL Lead department: Statistical Science
Lead Supervisor: Anna Calissano
Project Summary:
Why this research is important
Understanding cause-and-effect relationships from data is a central challenge across science, medicine, and technology. Causal discovery methods aim to identify how variables influence one another rather than simply describing their correlations. Graphs for Causal Discovery (Graphs4Causal) focuses on developing new statistical methods for sets of causal graphs. The project seeks to enhance the robustness and interpretability of causal graph estimation, leading to more accurate and reliable causal conclusions. The developed methodology will be applied to single-cell gene expression data from lung tissue, collected with and without exposure to steroids. The objective is to investigate the effects of steroids—commonly prescribed for severe asthma—on the gene expression profiles of lung cells, with the aim of tracing both their therapeutic and potential adverse effects.
Who you will be working with
The PhD will be supervised by Dr. Anna Calissano and Prof. Karla Diaz Ordaz at the Department of Statistical Science (UCL). The candidate will also interact with Dr. Rocio T Martinez-Nunez (KCL) and Dr. Irene Balelli (INRIA).
What you will be doing
The PhD project will focus on developing statistical methods for sets of causal graphs. Specifically, you will: Develop a robust estimator for causal graphs; Quantify uncertainty around estimated causal structures; Compare dimensionality reduction techniques applied at both the dataset and graph levels. The methods developed will be applied to single-cell gene expression data, which investigates the effects of steroid treatments for severe asthma on lung cells.
Who we are looking for
We are seeking a candidate with a degree in mathematics, statistics, computer science, or a related field. The ideal candidate will be motivated to work at the interface of theory and application, contributing to both methodological development and real-world data analysis within a collaborative, interdisciplinary research environment.