Severe Mental Illness PhD: Using Genetic and Single-Cell Data to Explore Metabolic Risk

Severe Mental Illness PhD: Using Genetic and Single-Cell Data to Explore Metabolic Risk

Quick Decision

This four-year MRC FGx-DLA project asks why some people with genetic liability to severe mental illness also develop obesity or type 2 diabetes, while others do not. The proposed sequence is to cluster schizophrenia and bipolar-disorder risk variants, test each cluster against metabolic and clinical outcomes in biobanks, integrate the results with Exeter single-cell expression and chromatin-accessibility data, then select candidates for functional follow-up. CRISPR perturbation and neuronal models are examples in the project materials, not confirmed core deliverables.

The project fits applicants interested in statistical genetics, psychiatric and metabolic disease, and functional genomics. Its public description does not identify the exact GWAS releases, biobanks, ancestry mix, access timetable or guaranteed experimental work. Exeter lists 5 January 2027 as the current deadline; an earlier vacancy record lists 8 December 2026. Confirm the operative date and, for international applicants, the written tuition treatment and 30% cohort cap before relying on the offer terms.

Research Problem and Doctoral Work

What remains unresolved

People with severe mental illness (SMI), particularly schizophrenia and bipolar disorder, experience elevated rates of obesity and T2D, yet metabolic outcomes vary substantially. The project asks why some people with high psychiatric genetic liability develop metabolic complications while others remain metabolically healthy. Its working hypothesis is that psychiatric risk is biologically heterogeneous: different variant sets may influence adiposity, glucose regulation or neither, through different pathways.

This question is difficult because a psychiatric diagnosis aggregates many genetic and clinical pathways. Observed comorbidity can also reflect medication, lifestyle, social exposures and healthcare patterns. Cross-disorder research finds both shared and diagnosis-specific common-variant influences, while a 2026 analysis of 249 circulating metabolites reports that major depression's genetic pattern differs from those for bipolar disorder and schizophrenia, despite more similar clinical overlap. These findings support analyzing diagnoses and mechanisms separately; they do not show that any one genetic pathway causes an individual's metabolic outcome. See cross-disorder psychiatric genetics and genetic overlap with metabolic markers.

Advertised work and methods

The project description lays out a linked sequence:

1. Genetic clustering: use large-scale GWAS data to identify genetically distinct clusters of psychiatric-risk variants and characterize each cluster's association with obesity and T2D.
2. Population consequences: examine biological and clinical correlates of the clusters in population biobanks.
3. Cellular interpretation: integrate genetic associations with Exeter single-cell multiomic data, including cell-type-specific gene expression and chromatin accessibility, to test whether metabolically distinct pathways implicate different neuronal or glial populations and to identify candidate genes and regulatory networks.
4. Functional prioritization: prioritize selected regulatory elements and candidate genes for possible functional work, such as CRISPR-based perturbation or neuronal cell models.

Statistical genetics, causal-inference methods and functional genomics are part of the advertised approach. The project also names R, Python, high-performance computing, transcriptomic integration and molecular/cellular laboratory skills as training areas. The description presents substantial room to shape the topic, including a particular SMI, ancestry-specific mechanisms, methods, neuronal populations or candidate-gene characterization. It does not specify the exact GWAS releases, sample sizes, ancestry balance, biobank outcomes, number of candidates for experimental validation or which assays are guaranteed. The sequence is a credible research plan, but the final breadth will depend on data access and the time needed to move from cluster discovery to functional work.

Tyrrell's 2026 MDD preprint is a close methodological precedent: MR-Clust grouped 621 MDD variants into three clusters with different, including opposing, associations with T2D and cardiometabolic profiles. This is encouraging evidence that the clustering idea can be operationalized, but the work is not yet peer reviewed and concerns MDD, not schizophrenia or bipolar disorder. A 2025 bidirectional Mendelian-randomization study of MDD and T2D likewise emphasizes pathway-specific results and BMI-related attenuation. Those studies motivate the question; they do not establish the answer for this project. See the 2026 clustering preprint and the 2025 MDD–T2D study.

The local data rationale is strengthened by a 2026 Exeter-affiliated single-nuclei resource that reports paired expression and chromatin-accessibility profiles across amygdala, hippocampus, hypothalamus and prefrontal cortex. It covers about 140,000 nuclei from 40 samples from 11 donors and integrates the resource with GWAS traits, including BMI, schizophrenia and bipolar disorder. The study is a preprint, and its donor count and post-mortem adult tissue limit what it can establish. The project description does not explicitly identify this publication as the exact dataset available to the student, so applicants should confirm which datasets are already generated, accessible and covered by approvals. See the multiomic resource.

Evidence-backed interpretation and limits

The research frontier is moving toward mechanistic stratification: large GWAS, cross-trait genetic models, cell-type-specific regulatory maps and follow-up experiments. A 2024 multi-ancestry T2D study identified eight genetic clusters across roughly 2.5 million participants and connected some signals to cell-specific chromatin profiles. That supports the general logic of pathway-based analysis, but it is a T2D classification and cannot be treated as an established clustering scheme for psychiatric risk. A cross-trait schizophrenia–T2D study combined GWAS and molecular-QTL evidence but found no genome-wide evidence of a direct causal relationship between the two disorders. Shared loci or cell-type enrichment therefore identify hypotheses, not clinical causation. See T2D genetic heterogeneity and schizophrenia–T2D genomic overlap.

An important analytic challenge will be separating psychiatric liability from direct metabolic genetic risk, medication and other environmental/clinical factors. In two UK SMI cohorts, physical-condition polygenic scores were more consistently associated with corresponding comorbidities than psychiatric scores, which were weaker and less consistent. A 17-million-person Scandinavian register study estimates that genetic factors contribute to about half of the comorbidity for some schizophrenia and affective-disorder/cardiometabolic pairs, while emphasizing that the balance differs by diagnosis and environment. These findings support a multi-cause model rather than a single genetic explanation. Potentially useful research questions include whether cluster associations remain after accounting for metabolic genetic pathways and treatment-related factors, and whether effects replicate across ancestry groups; these are analyst-suggested questions, not advertised tasks. See the UK SMI cohort study and the population-register analysis.

School, Programme and Field Context

The official project document names the University of Exeter as its primary institution and the Department of Clinical and Biomedical Sciences as its department. Exeter lists an MPhil/PhD in Clinical and Biomedical Sciences and offers full-time or part-time research study. The award itself is four years full time, with part-time study available. The university's PGR rules describe regular progress reviews, registration-linked access to facilities and supervision, and a maximum registration period determined by qualification and attendance mode. No joint or double degree is stated. The MRC FGx-DLA shares training and scientific activities across Exeter, Cambridge, Edinburgh, Imperial, King's and Oxford, but this network participation does not change the project document's named primary institution. See the Exeter research-degree page, Doctoral College enrolment guidance and MRC programme description.

For institutional context, the department reports that its Clinical Medicine research was in the UK's top 10 for world-leading and internationally excellent performance based on REF 2021, and gives Times Higher Education World University Rankings 2026 subject bands of top 125 for Life Sciences and top 175 for Medical and Health. These are subject or departmental signals, not a project-specific ranking or a guarantee of supervision quality. The department's stated themes include diabetes, molecular genetics, epidemiology, data science and precision medicine; its advertised facilities include single-cell omics, long-read sequencing, imaging and flow cytometry. Whether each facility is available to this student should be verified. See departmental research and facilities and Exeter's REF 2021 summary.

Supervisor and Lab Assessment

The project PDF names Jess Tyrrell as lead supervisor and provides j.tyrrell@exeter.ac.uk for project enquiries. Exeter describes her as an Associate Professor in Genetic Epidemiology whose research applies genetic and epidemiological methods to the relationship between metabolic and mental health. The fit between her recent work and the advertised clustering question is unusually direct. The PDF names no additional project-specific supervisors; confirm the supervisory team and the division between population genetics, single-cell analysis and experimental work. See the department leadership profile.

Three recent contributions frame the opportunity:

• Handley et al. (2026 preprint) use MR-Clust on MDD variants and report three clusters with distinct cardiometabolic profiles, including opposing associations with T2D. This is the closest precedent, but it is an unreviewed preprint and its findings cannot be carried over to schizophrenia or bipolar disorder.
• Bala et al. (2025, Molecular Psychiatry) examine bidirectional MDD–T2D relationships and identify pathway-specific associations, including body-fat and obesity-mediated insulin-resistance components. The paper supports the need to separate mechanisms; its MR estimates are not treatment effects and depend on instrument assumptions.
• Gebretsadik et al. (2026) test eight mechanistically partitioned T2D polygenic scores against psychiatric and related phenotypes in UK Biobank. Associations vary by T2D pathway, but the cohorts are White British/European and the analyses are not causal proof.

An Exeter-affiliated 2026 preprint provides a relevant cell-resolution precedent. It combines gene expression and chromatin accessibility in four post-mortem brain regions with GWAS, highlighting candidate neural populations for BMI, schizophrenia and bipolar disorder. Its reported T2D signal was concentrated in an obesity-associated cluster, with inhibitory neurons and oligodendrocyte precursors enriched. Because it is a small-donor resource study and currently a preprint, it is best read as a way to prioritize hypotheses, not as proof that a given cell population mediates clinical comorbidity. The project's exact overlap with this resource and access to any newer datasets remain to be confirmed.

Funding, Application and Living Cost

Award and date

The corrected award is the MRC Functional Genomics Doctoral Landscape Award (MRC FGx-DLA), reference 6025. The current Exeter programme page states that applications opened 26 October and close on Tuesday 5 January 2027. The captured advert records 8 December 2026. Because the earlier date remains unaccounted for, prepare to meet 8 December unless the admissions team confirms otherwise in writing, and ask for the exact time and time zone.

The official award lists four years full time (part time also available), UK tuition fees, a UKRI-minimum stipend of £21,805 per year for 2026/27, updated annually, and up to £5,000 per year of research-training and support funding. UK and international candidates may apply, subject to the programme's 30% international-student cap. UKRI grants cover only the Home fee level; however, the current national FGx-DLA page says Exeter students do not need to supplement UKRI funding. Because award 6025 states UK tuition, international candidates should obtain written confirmation of the fee arrangement and how the cap applies. See the MRC DLA FAQ, national FGx-DLA international-fee guidance, UKRI training-grant terms and UKRI stipend guidance.

Eligibility and application steps

The programme requires at least a 2:1 honours undergraduate degree or international equivalent in a relevant field. Listed fields include genetics, molecular biology, biomedical science, human biology, biochemistry, medicine, healthcare professions, bioinformatics, computer science and statistics. International applicants typically need IELTS 7.0 overall with no component below 6.5, or an accepted equivalent, evidenced by the deadline. The university portal application and the FGx-DLA applicant questionnaire are both required; only one application per person is accepted, with up to three project preferences listed in the questionnaire. The questionnaire asks about motivation and longer-term plans, an achievement/experience, research experience and reflection, and GenAI use (or non-use). Two references are required and must not be from the prospective supervisory team. If shortlisted, applicants must arrange for academic referees to email references seven days before interview; the team will not request them. Application documents must be in English, with certified translated copies of qualifications. Check the current page for form-specific instructions and ask about interview dates before planning references.

Exeter budget and career relevance

Exeter estimates basic student spending at £1,300–£2,000 per month, based on local spend in April 2026. Dividing the 2026/27 stipend by 12 gives an indicative £1,817 per month: roughly £517 above the estimate's lower end but £183 below its upper end, before one-off arrival costs. The university's postgraduate accommodation page links 2026/27 rents and room types; rent, contract length, bills and personal circumstances will determine the actual margin. See Exeter's budget estimate and postgraduate accommodation information.

The skills profile maps to academic genetics/epidemiology, research bioinformatics, functional genomics, biotechnology and precision-medicine research. Skills England estimates 66,000 additional jobs in selected life-sciences priority occupations between 2025 and 2035, plus replacement demand, and identifies pressure in data/digital and scientific roles. These are sector-wide, uncertain estimates and do not guarantee jobs for PhD graduates. NHS clinical bioinformatics is one adjacent route where biological knowledge, programming and data analysis intersect, but this doctorate alone would not establish clinical registration or specialist eligibility. See the Skills England sector assessment and NHS clinical-bioinformatics role profile. The wider MRC DLA includes cross-sector training and placements, but a placement for this particular award should not be assumed without confirmation; see UKRI's DLA overview.

Applicant Fit and Preparation

Required conditions

The published entry conditions are a 2:1 or equivalent in a relevant area and, where applicable, the English-language evidence above by the deadline. These are separate from the project's technical training list. Confirm degree equivalency, any English-test waiver and the date by which proof must be submitted.

Useful experience and preparation

The project spans statistics/genetics, population phenotypes, single-cell regulatory data and wet-lab validation. Applicants can strengthen an application by showing reproducible R or Python work, genetics/statistics or epidemiology experience, handling of large data, transcriptomic or single-cell analysis, or molecular/cellular biology. These are evidence of readiness, not additional formally stated entry prerequisites. A strong application should explain how a previous research task was designed, what its limits were and how the applicant responded to uncertainty. The questionnaire explicitly invites research reflection and discussion of the applicant's own use of generative AI.

This opportunity may be less suitable for candidates seeking only a clinical practice/psychology project, or for those unwilling to work across both data-intensive analysis and experimental validation. The degree of balance between computation and laboratory work is not specified, so ask which stages will be core requirements versus optional extensions.

Next Actions and Questions to Verify

Before applying

1. Ask Exeter admissions and Jess Tyrrell to confirm whether the deadline is 8 December 2026 or 5 January 2027, including the precise time zone and whether the two-stage process has a separate project-specific cut-off.
2. For international applicants, request written confirmation that the Exeter no-supplement policy applies to this award and whether the 30% cap is applied at programme or local-award level.
3. Confirm the applicant's degree/subject equivalence, IELTS evidence date, required documents, questionnaire link and reference format.
4. Ask whether the applicant may list three preferences and whether only one application should be submitted through Exeter; the current programme page says one application per person.

At interview

1. Which psychiatric GWAS releases, ancestries, phenotypes and biobanks will anchor the clustering work? How will sample overlap, medication and direct metabolic genetic liability be handled?
2. What is the status, access pathway and donor/ancestry composition of the Exeter single-cell multiomic data? Is the published four-region resource the dataset planned for this project?
3. How many candidates are expected to progress to CRISPR or neuronal models, which models/assays are already available, and what are the criteria for choosing them?
4. Who will supervise the statistical genetics, single-cell analysis and functional work, and how will shared FGx-DLA training be integrated with this project?

Before accepting an offer

1. Check the written award letter for the full duration, annual stipend schedule, UK tuition coverage, Exeter’s no-supplement fee arrangement and the 30% cap, and the exact uses/availability of the £5,000 annual support fund.
2. Confirm that the planned data access, compute allocation, laboratory access and supervisor time are in place for the advertised four-year sequence.

Evidence and Sources

Project, programme and institutional context

• Official project PDF
• MRC FGx-DLA programme page
• MRC FGx-DLA award record 6025
• Clinical and Biomedical Sciences PhD and Research Degrees
• Clinical and Biomedical Sciences research
• Department leadership and Jess Tyrrell profile
• Exeter REF 2021 results
• Doctoral College enrolment guidance
• UKRI overview of MRC DLA recruitment

Supervisor, psychiatric genetics and multiomics

• MDD genetic-clustering preprint
• Bidirectional MDD–T2D Mendelian-randomization study
• Mechanistically partitioned T2D polygenic risk and psychiatric phenotypes
• Exeter-affiliated single-nuclei multiomics resource
• Genetic overlap between T2D and schizophrenia
• Genetic overlap between severe mental disorders and metabolic markers
• Genetic architecture across 14 psychiatric disorders
• Genetic drivers of T2D pathophysiology
• Physical-health genetic liability in UK SMI cohorts
• Exon-level multi-modal analysis of severe psychiatric disorders
• Mental–cardiometabolic comorbidity in 17 million Scandinavians

Funding, cost and career evidence

• National FGx-DLA international fee guidance
• MRC DLA webinar FAQs
• UKRI standard terms for training grants
• UKRI stipend and fee guidance
• Exeter student budget estimate
• Exeter postgraduate accommodation information
• Skills England life-sciences assessment
• NHS clinical bioinformatics (genomics) role profile