Research Opportunities

How do neural circuits enable us to anticipate future events, learn from rewards, and acquire new skills? The Wagner lab investigates the circuit computations underlying cerebellar contributions to learning and cognition, with a particular focus on interactions between the cerebellum and neocortex. We ask how these interconnected brain regions transform experience into predictions and learned behaviors, and how they support generalization across experiences while limiting interference between them. Our recent publications in Nature (2026), Nature Neuroscience (2026), and Neuron (2024) address these questions through studies of contextual learning, reward prediction, and learned timing.

We combine multi-site two- and three-photon imaging, holographic photostimulation, and electrophysiology during novel skill learning in mice. These approaches allow us to follow identified neural populations across learning, observe communication between brain regions, and test how specific patterns of neural activity contribute to behavior. We develop behavioral tasks using robotic interfaces and virtual reality, integrating experiments with neural population analysis and computational circuit models.

For OxCam, participation in our lab requires a computational or theoretical neuroscience co-mentor at Oxford or Cambridge. Students will combine experimental research at NIH with complementary computational research in the UK, addressing a shared scientific question through models that guide experiments and experimental results that constrain theory. The UK co-mentor and specific research direction will be developed jointly with the student according to their interests and the scientific fit.

Interested applicants should contact Mark Wagner at mark.wagner@nih.gov.

Ocular Stem Cell and Transnational Research Section at the National Eye Institute (NEI), NIH, is focused on understanding the mechanisms of human retinal diseases and developing innovative therapeutic approaches to prevent vision loss and restore sight. Our research integrates stem cell biology, tissue engineering, artificial intelligence, functional genomics, and regenerative medicine to bridge fundamental discoveries in human biology with clinical translation. Using human induced pluripotent stem cells (iPSCs), we develop physiologically relevant models of retinal tissues to investigate disease mechanisms, identify therapeutic targets, and advance precision medicine approaches for age-related macular degeneration (AMD) and other inherited and acquired retinal degenerative diseases.

A major focus of our laboratory is developing iPSC-derived retinal pigment epithelium (RPE) tissue patches for transplantation into patients with advanced AMD. This program has progressed from fundamental stem cell research and tissue engineering to clinical testing, establishing a translational framework for developing regenerative therapies. Unlike conventional pharmaceuticals, transplanted cells represent living drugs that can persist long-term, adapt to their environment, and dynamically interact with surrounding host tissues. We seek to understand how the unique pathological environment of each patient influences the survival, integration, and function of transplanted cells, and how these cells, in turn, modify the recipient tissue, providing a new paradigm for precision regenerative medicine. Using the same principle, we are currently developing a photoreceptor cell therapy for retinitis pigmentosa and AMD. 

In parallel, we use patient-derived iPSCs and engineered human retinal tissues to recreate key aspects of retinal degeneration in vitro. Our research investigates how genetic susceptibility, aging, complement activation, lipid dysregulation, oxidative stress, and mitochondrial dysfunction converge to drive cellular degeneration. By combining these disease models with CRISPR-based functional genomics, high-content imaging, and molecular profiling, we aim to identify causal disease mechanisms and discover therapeutic interventions capable of restoring cellular homeostasis. Our work has led to repurposing of drugs like metformin and pioglitazone for AMD and monogenic retinal degenerative diseases such as ciliopathies. Currently, we are in the process of developing gene agnostic gene therapies for these conditions. 

We are also pioneering the integration of artificial intelligence and cellular digital twins into human disease modeling. By combining high-resolution three-dimensional imaging, machine learning, and quantitative analysis of subcellular architecture, we generate computational representations of human RPE cells that reveal how organelle organization, cellular polarity, and tissue architecture change during development, aging, and disease. These approaches enable the discovery of previously unrecognized cellular biomarkers and provide a foundation for predicting disease-associated cellular states and responses to therapeutic interventions.

Together, these interconnected research programs establish a comprehensive framework linking human disease modeling, mechanistic discovery, computational prediction, and clinical translation. Our overarching vision is to transform regenerative medicine from a one-size-fits-all approach into a predictive, personalized therapeutic strategy, in which the unique biology of each patient informs the development, selection, and evaluation of treatments capable of preserving and restoring vision.

Artificial intelligence (AI) is transforming biomedical research and healthcare by enabling new approaches to analyzing complex biomedical data, generating scientific knowledge, and supporting clinical decision-making. Our laboratory develops and evaluates cutting-edge AI methods, including large language models (LLMs), foundation models, natural language processing, and multimodal AI, to address important challenges in biomedicine.

Potential research directions include AI-driven scientific discovery, biomedical knowledge extraction and integration, AI agents for complex research tasks, multimodal learning from biomedical text and images, and AI applications in precision medicine and clinical research. We are also interested in developing reliable, interpretable, and evidence-grounded AI systems that can be effectively applied to real-world biomedical problems.

Specific doctoral research topics will be tailored to the student's interests, expertise, and the complementary strengths of the NIH and Oxford/Cambridge laboratories. Projects may involve novel AI methodology development, integration of heterogeneous biomedical data, rigorous model evaluation, and interdisciplinary collaborations with biomedical and clinical researchers.

We welcome highly motivated students with backgrounds in computer science, artificial intelligence, computational biology, biomedical informatics, or related disciplines who are interested in advancing AI methods and their applications to biomedical discovery and human health. More lab information can be found at https://irp.nih.gov/pi/zhiyong-lu

GLP-1 receptor agonists (GLP-1R agonists or GLP-1RAs) are a class of drugs that mimic the human GLP-1 hormone to manage type 2 diabetes and obesity, that are being taken by millions of Americans. However, not all people respond to GLP-1RAs. One explanation is genetic variation in GLP-1R. Over 500 genetic variants of GLP-1R have been described in the human population. This project aims to characterize how these variant GLP-1R proteins vary in their sensitivity to GLP-1RAs.

Some GLP-1RA drugs are based on a naturally occurring component of Gila monster venom, exendin-4. Gila monsters have two other homologs of GLP-1. Evidence suggests that exendin-4 may have evolved as a defensive peptide in Gila monsters against predators (largely mammals). The specificity and longevity of the Gila monster exendin-4 against the variation that exists across vertebrate species is still an open question, and will also be characterized in this project. 

This project will investigate how chromosomal instability (CIN) shapes systemic and tumour-associated myeloid responses in oesophagogastric cancer (OGC), with a particular focus on neutrophil-driven immunosuppression as a therapeutic target. Building on the hypothesis that CIN promotes a targetable, immunosuppressive myeloid infiltrate, the work will combine longitudinal patient sampling with high-dimensional immune profiling to characterise circulating and tumour-resident myeloid populations at baseline and during chemo-immunotherapy. Using matched tumour and blood samples, CIN status will be quantified via cGAS-positive micronuclei, and correlated with dynamic changes in circulating immune cells and chemokine profiles to identify biomarkers of treatment response and resistance. Parallel analyses in novel CIN-high and CIN-low mouse models will enable mechanistic dissection and testing of strategies to therapeutically target myeloid-mediated immunosuppression. Together, this work aims to define clinically actionable biomarkers and vulnerabilities in CIN-high cancers, enabling earlier patient stratification and more effective, personalised treatment approaches.

Despite the improvement in cancer treatment achieved through immune checkpoint blockade (ICB), durable curative responses are realised in only a minority of patients with certain cancers. This is best evidenced in colorectal cancer (CRC), where only the small subset of patients with deficient mismatch repair disease receive real benefit from ICB. A fundamental challenge for most CRC patients is the high prevalence of liver metastases. This is characteristic of late-stage disease and is strongly associated with poor response to treatment and survival. The overarching aim of this project is to better understand key suppressive mechanisms in primary and metastatic tumours, insight that is vital to support the design of better treatments combinations tailored for different CRC patients.

To interrogate the immunosuppressive mechanisms that form in CRC, we have developed sophisticated orthotopic models utilising tumour organoid implantation directly into the colon wall via colonoscopy. Slow growing primary tumours form and then naturally metastasise to the liver. Exploiting dynamic labelling approaches pioneered with the lab, mechanistic studies focus on understanding how different tumour microenvironments shape immune cell fate and function in real time. These studies are then further supported by conditional knockout mouse models and other interventions (e.g. blocking antibodies) to test effects on anti-tumour responses.

This project will provide a wealth of training opportunities and is ideal for students wishing to develop expertise in studying immune responses to cancer utilising the most advanced in vivo models available. The lab is medium-sized (~10 members), with dedicated technical support for in vivo research and a collaborative and supportive ethos.

The shape, spatial organisation, and communication pathways of subcellular organelles are central to neuronal function and are profoundly disrupted in neurodegenerative diseases such as ALS, Alzheimer’s, and Parkinson’s disease. This PhD project will combine advanced live-cell and super-resolution microscopy, multiomic profiling, and CRISPR-based gene editing in human iPSC-derived neurons to define how organelle form supports neuronal function, and how these relationships break down in neurodegeneration. The student will develop and apply quantitative imaging approaches to map organelle architecture and interactions, integrate these data with proteomic and lipidomic analyses, and functionally perturb candidate pathways to establish causal links between organelle organisation and neuronal dysfunction.

Advances in single-cell transcriptomic technologies are enabling the discovery of many novel cell phenotypes. However, this emerging knowledge remains fragmented across the scientific literature. Natural language processing (NLP) using large language models (LLMs) and other artificial intelligence methods offers a promising approach to extract and organize this information at scale, but the inconsistent nomenclature used to describe cell phenotypes in the literature limits the effectiveness of straightforward NLP approaches, and more advanced NLP methods requiring well-annotated corpora for development and evaluation are needed. We recently developed the NLM CellLink corpus, a corpus of excerpts from full-text articles that contain information about cell phenotypes. This corpus was manually annotated with mentions of human and mouse cell types and linked to Cell Ontology (CL) identifiers. This PhD project will use this corpus to support the development and evaluation of AI machine learning models for automatically identifying cell types in the scientific literature, including novel cell types, and their relationships with other key biological entities (e.g., marker genes, anatomical structures, perturbation responses, disease states) for translation into standardized semantically structured assertions and their incorporation into the NLM Cell Knowledge Network (NLM-CKN).

The student will join an interdisciplinary team at the National Library of Medicine that is applying computational biology and data science techniques to characterize cellular phenotypes and their roles in health and disease at scale.  Training will be provided in advanced computational and statistical analysis of multi-omics data, natural language processing using artificial intelligence methods, and the development and use of ontologies and other sematic web technology for biomedical knowledge representation.

Cells are the fundamental units of life. Single cell genomic technologies are revolutionizing our understanding of cellular phenotypes. Large single cell data consortia, including the NIH BRAIN Initiative and the Human BioMolecular Atlas Program (HuBMAP), have generated single cell atlas data from millions of cells/nuclei spanning multiple organs and biological systems. At the National Library of Medicine (NLM), we are building the NLM Cell Knowledge Network (http://cell-kn-mvp.org), a knowledgebase that focuses on representing the cell phenotypes and associated characteristics derived from single cell genomics data. It integrates data-driven information with knowledge from trustworthy reference ontologies, NCBI resources, and text mining efforts, resulting in a large-scale semantic knowledge network for innovative data mining and knowledge discovery.

This project consists of two main research components: 
i) developing novel computational methods for single cell and spatial transcriptomics analysis using machine learning and advanced statistics techniques, and 
ii) developing network analysis strategies for knowledge mining using cutting-edge artificial intelligence technologies. 

Students interested in one or both research components are encouraged to apply. The project team has interdisciplinary background, ranging from molecular biology, genetics, statistics, and computer science, providing a strong supporting system for students’ academic growth. Dr. Yun (Renee) Zhang is a tenure-track investigator at NLM and an alumnus of the University of Oxford
 

Sympathetic neurons have a wide range of physiological functions and their hypoactivity contributes to obesity and diabetes, among other syndromes. Sympathomimemic drugs rescue this deficiency but this drug class, mostly composed of brain-penetrant amphetamines and adrenergic agonists, is both cardiotoxic and highly controlled. Our recent publication puts forward new class of drugs  named Sympathofacilitators that do not enter the brain and have an anti-obesity and cardio-neutral effect in vivo. The first-in-class was published in Mahu I et al Domingos, Cell Metabolism 2020; Fig. 3C of this paper demonstrated a neuro-facilitatory effect, rather than neuro-excitatory one. 

This new class is in needed of novel chemical entities which can be screened in vitro on cultured iPSC-derived sympathetic neurones. The screen would be based on fluorescent readouts of calcium activity reporter, screening for a facilitation of responses to acetylcholine (similar to Fig. 3C of Mahu I et al). 

The prospect of identifying natural compounds that have a Sympathofacilitatory effect is tangible when performed in collaboration with the laboratory of Barry O’Keefe. The student will learn lab how to grow and scale-up iPSC-derived sympathetic neurones in Domingos lab, and optimize an in vitro assay based on Fig. 3C. The student will then transfer this knowledge to the lab of Barry O’Keefe where the screen will be performed using a fluorescent plate reader, robotic liquid handling, and a library of natural compounds.

While X-ray crystal structures of proteins are terrifically informative, both in basic research (6 Nobel prizes in 25 years) and drug design, the major stumbling block remains the need to obtain crystals of the biological sample (protein, protein complex, virus) suitable for X-ray analysis.  Even for a biochemically well-behaved sample, the likelihood of it crystallizing is always low, and there are no rules or reliable protocols to improve this.  Structure-based drug discovery similarly requires diverse crystal forms to be available for the protein; this too cannot currently be easily induced.

This project addresses the pressing need for rapid and simple protein engineering techniques to make samples crystallize routinely.  This will entail, depending on the student's background, skills and preferences:

• Develop streamlined protocols for parallel generation of large numbers of crystallization chaperones, e.g. using in vivo selection directly coupled to over-expression, with DNA libraries of binding scaffolds;
• Evolve binders that favour crystallization, by adapting existing motifs through iterative design supported by high-throughput crystallization, e.g. Gluebodies [Ye et al, 2024]; 
• Adapt protein design algorithms to find surface mutations that better allow the protein to pack into crystals, e.g. by machine learning of crystal contacts in the PDB;
• Streamline protocols for rapid large-volume parallel expression, purification and crystallization of large numbers of diverse protein variants; 
• Develop generic scaffolds that crystallize robustly, and ways of tethering target proteins to the scaffold.

Cardiometabolic disease (CMD) remains the leading cause of global mortality, contributing to a heavy burden on healthcare systems. It is well accepted that lifestyle factors, such as poor diet, physical inactivity, smoking, and excessive alcohol consumption, are important contributors to CMD incidence and mortality but underlying biological mechanisms linking lifestyle factors with cardiometabolic health are largely unknown. Research on such mechanisms may uncover novel biomarkers that more accurately predict disease onset, progression, and response to lifestyle interventions. 

The China Kadoorie Biobank study (CKB) is a large prospective cohort study with >0.5 million Chinese adults from 10 different locations in China (www.ckbiobank.org). Over the past 20 years, CKB has accumulated multi-dimensional data on lifestyle and other exposures, physical and other measurements (e.g. adiposity, blood pressure, liver steatosis and fibrosis, ECG, carotid artery intima media thickness and plaque, bone mineral density, and retinal images), incidence and mortality of major diseases including CMD, as well as multi-omics data including genomics (GWAS genotyping as well as WGS), proteomics (>10,000 proteins from Olink and Somalogic), metabolomics (>220 NMR and >5400 Metabolon metabolites, which cover 8 super biological pathways, e.g. amino acid metabolism, nucleotide metabolism, and microbiome metabolism, and 70 major pathways) and gut/oral metagenomics (shotgun sequencing). This large and rich resource will enable us to investigate the potential relevance of different lifestyle factors, as well as their interplay, for a range of cardiometabolic health conditions. Findings from CKB could be compared with those from the UK Biobank, which is an open resource for global researchers.

Epigenetic regulation plays a crucial role in maintaining homeostasis by fine-tuning gene expression. As we age, epigenetic patterns become deregulated, contributing to age-related conditions such as senescence, metabolic disorders (e.g. diabetes, atherosclerosis), immune dysfunction, and cancer, which collectively decrease life expectancy and quality of life in the elderly. Histone deacetylases (HDACs), enzymes that remove acetyl groups from lysine residues thus promoting chromatin condensation, are key regulators of gene expression and their activity contributes to age-related deficiencies in homeostasis. HDAC inhibitors offer potential to reverse aging by restoring health-promoting and anti-senescence acetylation patterns, improving both cellular and systemic functions and alleviating the aging process. Our previous research demonstrated that HDAC inhibition has strong immunomodulatory effects in aging mice where it is able to revitalise the immune response, and potentially overcome immune related senescence. Specifically, enhanced MHC I and II expression, increased activity of immune cells and infiltration into the tumour microenvironment. This project will investigate the effects of clinical HDAC inhibitors on immune aging, with a focus on restoring healthy immune responses and protection against cancer and age-related diseases, both in vitro and in vivo. The DPhil student will be trained in cutting-edge molecular, biochemical, genomic, bioinformatic and immunological techniques including but not limited to ChIP-seq, single cell RNA seq, CRISPR/Cas9 gene editing, flow cytometry, and ELIspot.

Liu G, Barczak W, Lee LN, Shrestha A, Provine NM, Albayrak G, Zhu H, Hutchings C, Klenerman P, La Thangue NB. The HDAC inhibitor zabadinostat is a systemic regulator of adaptive immunity. Commun Biol. 2023 Jan 26;6(1):102.

Blaszczak W, Liu G, Zhu H, Barczak W, Shrestha A, Albayrak G, Zheng S, Kerr D, Samsonova A, La Thangue NB. Immune modulation underpins the anti-cancer activity of HDAC inhibitors. Mol Oncol. 2021 Dec;15(12):3280-3298.

This collaboration between the Altan-Bonnet (NCI), Buckley and Coles (Oxford) labs addresses how organs generate distinct inflammatory responses despite sharing common components like immune cells, fibroblasts, and the extracellular matrix. In collaboration with additional teams from the Netherlands, and Canada, we aim to uncover the molecular, cellular, and tissue-level rules governing organ-specific inflammation.

We hypothesize that (1) organ context and cellular experience shape the perception of inflammatory signals, and (2) organ-specific hierarchies integrate responses into coordinated outcomes. Using a data-driven approach, we will combine ex vivo and in silico models of mouse and human tissues to explore these mechanisms. High-throughput robotics will generate diverse tissue models with varying sensitivities to infection or immunopathology. Multimodal datasets from these models will be analyzed using machine learning to build computational models, to guide iterative cycles of discovery.

This project will revolutionize tissue biology by creating a unified framework for understanding tissue-specific inflammation, paving the way for new treatments. We are seeking researchers with expertise in bioengineering, computer science, or immunology to join this interdisciplinary effort.

Project keywords: immunology, systems biology, biomedical engineering. 

This collaboration between the Altan-Bonnet (NCI), Buckley and Coles (Oxford) labs addresses how organs generate distinct inflammatory responses despite sharing common components like immune cells, fibroblasts, and the extracellular matrix. In collaboration with additional teams from the Netherlands, and Canada, we aim to uncover the molecular, cellular, and tissue-level rules governing organ-specific inflammation.

We hypothesize that (1) organ context and cellular experience shape the perception of inflammatory signals, and (2) organ-specific hierarchies integrate responses into coordinated outcomes. Using a data-driven approach, we will combine ex vivo and in silico models of mouse and human tissues to explore these mechanisms. High-throughput robotics will generate diverse tissue models with varying sensitivities to infection or immunopathology. Multimodal datasets from these models will be analyzed using machine learning to build computational models, to guide iterative cycles of discovery.

This project will revolutionize tissue biology by creating a unified framework for understanding tissue-specific inflammation, paving the way for new treatments. We are seeking researchers with expertise in bioengineering, computer science, or immunology to join this interdisciplinary effort.

Project keywords: immunology, systems biology, biomedical engineering.