St Andrews PhD: Multi-omics and AI for Brain Health and Ageing

St Andrews PhD: Multi-omics and AI for Brain Health and Ageing

Research Question and Fit

This is a four-year, competitively funded doctorate using existing cohort data to investigate genetic contributions to brain health and healthy ageing. The scientific challenge is to distinguish a useful prediction from an explanation of brain-health trajectories. Training combines neurogenetics, imaging AI and exposure to Scottish Brain Sciences (SBS). Bioinformatics, mathematical modelling and AI experience are advantages, explicitly not prerequisites. The opportunity suits applicants interested in careful population-data interpretation rather than generating new experimental data or delivering a treatment.

The School announcement lists 16 November 2026, without a clock time or time zone. The project’s funding terms provide UKRI maintenance, Home tuition and a St Andrews-funded international fee difference. The main practical uncertainty is data readiness: UK Biobank’s access notice describes a pause in new applications and planned reopening in late 2026; the subsequent status and this student’s permission remain unconfirmed. Confirm the student's approved access route and a workable fallback with the supervisor before relying on immediate analysis. Current project.

Research Problem and Doctoral Work

The proposed work combines genomic, proteomic, imaging, cognitive and demographic information from epidemiological cohorts such as UK Biobank. It asks how genetic associations with neurodevelopmental or neurodegenerative traits relate to brain-health trajectories. Existing data reduce the need to recruit a new large cohort; they do not establish that every proposed measurement exists in the same participants, at comparable ages or repeatedly over time. The first useful task is therefore to establish the available sample overlap, time structure and permissions before choosing an ambitious integration model.

Scientifically, predictive accuracy and an explanation of ageing are different outcomes. Age, education, health, ancestry and recruitment patterns can affect both measured traits and apparent genetic associations. Missing data and selective participation can change which population a result describes. These are analytical concerns derived from the advertised design, not additional compulsory admission requirements. A defensible thesis could compare interpretable baselines with integrated models, explain how much information each data type adds, and assess whether findings generalise across relevant populations or collection periods.

High-capacity models also create opportunities for misleading validation: measurements from the same person or information unavailable at the intended prediction time must not quietly enter a supposedly independent evaluation. The appropriate validation unit and target population should be agreed before interpreting performance. Neither a genetic association nor an improved classification score establishes that modifying a factor prevents dementia. The advertised contribution is an analytical framework, with causality and clinical use requiring additional evidence.

Access is a substantive dependency. UK Biobank requires approved public-interest research, organisational agreements and authorised researchers; restricted UKB-RAP access is the default. UK Biobank’s access notice describes a pause in new applications and intended reopening in late 2026; that plan does not establish whether data will be available to this project on its timeline. Paracchini's public project 80652 is marked Closed; that is evidence about one historical project, not proof that all her current access is absent. Ask which active approval covers this doctorate, which data are ready, and who pays analysis costs. Access rules, Current pause, Historical project.

School, Programme and Field Context

The awarding institution is the University of St Andrews, Scotland, United Kingdom; the host is its School of Medicine, with a Computer Science co-supervisor. The specified application course is PhD with internship. SBS is a company and placement collaborator, while GenLang is a consortium; neither is an additional degree-awarding institution. This is a research doctorate, not the clinical MD route or a qualification to practise medicine. Students are not university staff merely because they receive a studentship. Medicine research environment, School supervision policy.

The project announcement funds four years. General School policy describes a four-year submission period including continuation, at least two supervisors, at least ten recorded meetings annually and annual progress review, normally by month nine. These institutional requirements do not establish weekly individual supervision or automatic funding for extensions. Confirm the internship programme's actual registration and award calendar rather than equating a general continuation rule with a funded extension.

The School reports over 86% of its research as world-leading or internationally excellent in REF 2021, published in 2022. This is historical School-level research evidence, not a current overall university ranking, an AI-specific result or a guarantee about supervision. School REF evidence.

Training joins genetics, AI and neuroimaging. The advert offers SBS experience of cognitive assessment, imaging, data collection and governance, alongside consortium collaboration. The placement's allocation, timing and expenses need confirmation; it is not an employment offer. University GRADskills and dedicated PGR career support cover academic and non-academic planning. Health-data analysis, research software and data stewardship are plausible transferable directions, rather than measured employment outcomes for this project. SBS, GenLang, Doctoral training and careers.

Supervisor and Lab Assessment

Silvia Paracchini, the principal supervisor, is Professor of Neurogenetics and Genomics, leading research on complex cognitive and behavioural traits. Her public contact is sp58@st-andrews.ac.uk. Co-supervisor David Harris-Birtill is a Senior Lecturer in Computer Science, contact dcchb@st-andrews.ac.uk. The current project confirms regular meetings and reports but leaves frequency and the division of scientific, computational and day-to-day support unspecified. Paracchini profile, Harris-Birtill profile.

The three studies below address model fusion, developmental genetic association and imaging evaluation. None establishes that integrating these data already explains brain ageing.

Model fusion and robustness: BioFuse

Hossain and Harris-Birtill, PLOS One, 2026, DOI 10.1371/journal.pone.0320989: BioFuse combines representations from pretrained biomedical models and evaluates downstream classification. It reported the highest AUC on five of twelve benchmark datasets. The study shows why fusion is not automatically better: clean-data gains can coexist with worse corruption robustness, while genomic and longitudinal outcomes and demographic fairness require further evaluation. Its relevance is model comparison and validation; it does not validate this proposed brain-ageing system. Original article.

Genetic association across development: childhood grip strength

Abbondanza et al., Genes, Brain and Behavior, 2024, DOI 10.1111/gbb.70003, with Paracchini: childhood grip-strength analyses use 5,450 ALSPAC and 1,162 Raine participants to examine age-related context in genetic associations. The authors argue that shorter environmental exposure may aid interpretation, while reported polygenic-score associations did not survive multiple-testing correction. This supports developmental epidemiology expertise, not a demonstrated mechanism of brain ageing. Institutional paper record.

Independent imaging evaluation: cervical histopathology

Mohammadi et al., PLOS Digital Health, 2024, DOI 10.1371/journal.pdig.0000381, with Harris-Birtill: an imaging-classification study used 1,738 slides for training/validation and 811 independent test slides, reporting 93.4% malignant sensitivity. The study identifies annotation limitations, class imbalance and overlapping categories, and calls for larger-scale validation. These are useful lessons about evaluation and labels. Cervical histopathology is a different task from brain ageing, and this result is not a clinical-performance promise for the doctorate. Original article.

These are genuine supervisor-coauthored originals, not three completed brain-ageing demonstrations. Data permissions, compute allocation and statistical support still need project-specific confirmation.

Funding, Application, and Living Cost

The exact funding note supports a complete international competition route: four years, maintenance at the UKRI rate, Home fees and a university-funded international difference. Current EASTBIO FAQ corroborates international eligibility and fee-waiver arrangements. Its 30% international annual-intake cap is a network constraint, not a Chinese exclusion or an individual probability. Research grants and placement expenses support study; they are not personal disposable income. The individual award outcome remains subject to selection. Current funding FAQ.

The £21,805 UKRI minimum is for 2026/27, not a confirmed 2027 personal rate. The current residence table lists self-catered standard Fife Park at £8,847 for 50 weeks and en-suite Powell Hall at £12,739 for 50 weeks; most residence packages include utilities and internet, with exceptions. These are current table references, not guaranteed 2027 offers or a complete yearly budget. Against the 2026/27 stipend, the Fife Park example leaves £12,958 before the other two weeks, food, transport and all other costs. Availability, family accommodation and inflation can materially change this illustration. UKRI rates, Residence costs.

EASTBIO excludes visa and health-surcharge expenses. The cited charges are £558 for a Student visa and student IHS £776 per visa year, with duration and part-year rules; the visa can outlast the course. Budget relocation, deposits and family costs separately, and confirm written award dates and continuation liabilities. Student visa, IHS.

Applicant Fit and Preparation

EASTBIO requires a UK 2:1 undergraduate degree or international equivalent; a master's is not mandatory. A lower undergraduate result with a master's or significant relevant professional experience may receive individual assessment. The China guidance provides an affirmative route through an accredited bachelor's degree, with a 78% general minimum dependent on institution and degree, while directing Medicine applicants to course-specific requirements. It does not guarantee that every 78% applicant meets this particular PhD's equivalence. The MD's clinical-qualification rule should not be transferred to this research doctorate. China guidance, Research entry requirements.

Bioinformatics, mathematical modelling and AI are preferred experience, not mandatory prior mastery. Evidence of quantitative reasoning, relevant biological understanding and learning across disciplines is useful. A short account of a previous analysis should explain its question, validation and limitations; this is suggested preparation, not an additional mandatory project document. The opportunity is less suitable for someone seeking guaranteed new-data generation, direct clinical practice or an already defined intervention.

The University sets Profile 7-D as its general PGR minimum, but Schools normally require their taught-programme scores or higher. Medicine is listed at Profile 3-D in that table; its application to this specific research doctorate should be confirmed with the School. Its current table lists IELTS 7.0, each 6.5, including One Skill Retake; pre-2026 TOEFL 91, Reading 19/Writing 23/Listening 19/Speaking 22, and 2026-onward 5.0, each 4.5. Tests are normally within two years of the start. An English-taught degree from another country may be considered through verified English-medium study plus a current QS-top-300 institution, or ECCTIS C1 verification; a Chinese English-medium degree is not automatically exempt. The University allows missing English evidence to become an offer condition, while EASTBIO warns that evidence may be needed before a formal offer; confirm the host’s timing rather than treating a test certificate as universally mandatory at initial submission. English policy, Profile 3-D.

Submit through the St Andrews portal, select PhD with internship, complete the anonymous EDI survey and EASTBIO form, and provide qualifications, English evidence if applicable and two EASTBIO-template references. Referees may upload using the automated invitation or email Rachel Horn, pgrecbiology@st-andrews.ac.uk. Current EASTBIO permits one project across all hosts and requires references by 16 November; there is no published two-week grace here. Referees must not belong to the proposed supervisory team. If a reference may be delayed, contact Rachel: a host may consider an extension, but this is not an automatic grace period. Non-English academic documents require both the original and an official English translation. Generic university CV/proposal instructions and EASTBIO's restricted assessed-document list should be reconciled with Rachel rather than assuming additional assessed materials. University route, EASTBIO documents.

Career and Transferable Research Skills

A thesis that explains cohort overlap, validation choices and uncertainty could demonstrate skills in health-data analysis, research software and data stewardship. SBS experience can help connect cognitive and imaging data with governance requirements. Neither the placement nor model-development training confers clinical authority or guarantees employment. Discuss how to demonstrate individual contributions when data permissions restrict sharing. The project has no published employment outcomes.

Next Actions and Questions to Verify

Before applying, discuss the question with Paracchini and choose the single EASTBIO project. Arrange template references and translations early; confirm the portal's exact programme label, Medicine English profile and any additional university documents. Submit by 16 November 2026; neither the project nor current central instructions supplies a clock time or time zone. The current network selection process assesses both project suitability and interdisciplinary programme fit. Selection process.

At interview, prioritise the actual approved dataset, overlap between modalities, repeated measurements, validation plan, computing budget and a fallback if UK Biobank access remains delayed. Confirm how the two supervisors share work, meeting arrangements and SBS placement support. Existing datasets and past permissions do not settle these questions.

Before accepting, obtain written confirmation of the full international fee difference, four-year maintenance calendar, 2027 rate, research and placement expenses, and responsibility for continuation. Establish the actual start date: the source gives a deadline and four-year funding, not a project-specific start day. Check registration, English conditions, CAS and any offer-specific ATAS requirement. The funding framework is credible; data readiness remains a practical question requiring a project-specific answer.

Evidence and Sources

The cited papers provide distinct evidence on model robustness, developmental genetic association and imaging evaluation; none establishes that multi-omics integration has explained brain ageing. The project listing gives a deadline but does not confirm current application availability. The UK Biobank notice’s planned reopening does not establish project-specific data access. Dated fees and stipend rates are references, not confirmed terms for an individual award.

Project, School and Training

• Current School project announcement
• Scottish Brain Sciences: official company site
• GenLang consortium
• Medicine: REF 2021 research quality
• Graduate School training and PGR careers
• Medicine postgraduate research environment
• Medicine supervision and progress handbook

Supervisors, Papers and Data

• Silvia Paracchini: official profile
• David Harris-Birtill: official profile
• BioFuse: original 2026 article
• Childhood grip-strength GWAS: original abstract
• Cervical-biopsy AI: original 2024 article
• UK Biobank data access and governance
• UK Biobank: current new-access application status
• Paracchini historical UK Biobank project 80652

Funding and Application

• EASTBIO 2027 application instructions
• EASTBIO international funding and eligibility FAQ
• University research application route
• China postgraduate qualification guidance
• Postgraduate English policy and exemptions
• English Profile 3-D: current test table
• Current postgraduate residence fee table
• UKRI stipend reference by academic year
• UK Student visa and fees
• Student immigration health surcharge
• University research admission requirements
• EASTBIO current selection process