Building interpretable foundation models for predicting perturbation effects from high-throughput experiments (Professor Christopher Yau)
PROJECT TITLE
Building interpretable foundation models for predicting perturbation effects from high-throughput experiments
SUPERVISORS
DESCRIPTION OF PROJECT
Single-cell RNA sequencing and CRISPR screening enable high-throughput analysis of genetic perturbations at single-cell resolution. Understanding combinatorial perturbation effects is essential but challenging due to data sparsity and complex biological mechanisms. Our recent work has shown that scalable Gaussian Process models can be used to predict perturbation effects from such experiments (https://www.nature.com/articles/s41467-025-61165-7). This project aims to scale up or devise new methodology to handle much larger data sets such as the Tahoe-100M (https://huggingface.co/datasets/tahoebio/Tahoe-100M) and create a foundation model for predicting perturbation outcomes.
TRAINING OPPORTUNITIES
The project will enable the student to gain skills in state-of-the-art artificial intelligence research and modelling. Computing skills will be developed in the use of modern deep learning frameworks and large (bioinformatic) data handling and analysis.
Funding Information
The position is not currently funded and therefore the candidate will need to secure funding.
HOW TO APPLY
To apply for this research degree, please click here.