PhD Studentship: Trustworthy AI for Tissue Gene Expression

This four-year PhD studentship focuses on developing Blood2Tissue, an innovative AI modeling framework designed to predict tissue-specific gene expression programs directly from routine blood samples. Hosted at the Earlham Institute within the AI Integrated Digital Twins Group, the project addresses the challenge of making reliable inferences about inaccessible tissues like the brain or liver. The selected candidate will build upon existing cross-modal attention and uncertainty frameworks, utilizing large biological datasets and collaborating with external bioinformatics teams. The position offers comprehensive training in machine learning, transcriptomics, and high-performance computing within a vibrant research community.

• Design and calibrate the Blood2Tissue AI modeling framework to predict tissue gene expression from blood data
• Benchmark existing prediction methods using the GTEx dataset and reference cell atlases
• Investigate fundamental biological questions concerning blood-visible versus tissue-autonomous molecular programs
• Collaborate with external research teams such as the Open Targets team at the European Bioinformatics Institute
• Utilize high-performance computing resources and implement reproducible software development practices
• Engage in structured training programs spanning machine learning, single-cell genomics, and responsible AI