Edinburgh PhD in Open Science: Measuring Reuse of Shared Data, Models and Code

Edinburgh PhD in Open Science: Measuring Reuse of Shared Data, Models and Code

Research Question and Applicant Fit

This Edinburgh project studies the gap between making research outputs public and other people actually using them. It combines an institutional research-data-management setting with article/citation analysis and possible interviews to measure reuse and credit; actual repository and participant access still requires agreement. The combination suits data-science or quantitatively minded bioscience, environmental or social-science experience focused on research practice and attribution.

Project instructions specify EUCLID submission by midday (timezone unstated) on 16 November 2026 for 1 October 2027 entry and allow references up to two weeks later. Central EastBio guidance requires references by 16 November; clarify the conflict with the School in writing and plan for the earlier deadline unless an extension is confirmed. Funding is competitive. UK and international applicants, including those resident in mainland China, may compete, subject to academic and university entry conditions. The available material does not confirm whether this exact project is still accepting applications; confirm current status with the School and discuss research scope and data access with Andrew Millar.

Research Problem and Doctoral Work

Sharing is not the same as reuse

The core question is who reuses institutions' shared datasets, models and code, how they describe that use, and how the creators receive credit. The project extends existing assessments of data sharing to downstream behaviour. Biological, social and environmental sciences are explicitly within the comparison. AI models and data originating in electronic laboratory notebooks (ELNs) expand the scope beyond conventional supplementary datasets.

Four questions define the advertised work:

1. Compare the sharing of digitally created outputs, including ELN data and AI models, with existing assessments of data sharing.
2. Examine how datasets, models and code are mentioned in papers that share them and papers that reuse them, across the three disciplinary areas.
3. Match citations in the global Data Citation Corpus to datasets shared by participating institutions, treating citation as one scalable indicator of reuse.
4. Study how researchers describe reuse and credit. Surveys, focus groups and interviews might contribute; they are possibilities, not a fixed mandatory design.

The materials include in-house output repositories, global citation data and emerging metaresearch resources. Programmatic analysis, including LLM-based methods, is named. Work with BioRDM researchers and groups using collaborators' ResearchSpace RSpace ELN is part of the setting. Training covers data science and data curation. No specific programming language, LLM provider, compute allocation or lab assay is required by the project description.

The scientific difficulty is measurement validity. A data citation can acknowledge a resource without demonstrating that its data were analysed. Reuse can also occur without a formal citation. Code availability, successful execution, adaptation of a model and substantive scientific reuse are different outcomes. A defensible thesis should make these distinctions explicit and test the limitations of its indicators. DataCite documents a public corpus built from multiple citation sources; Crossref explains why informal references in article text can be missed by formal citation workflows. Corpus documentation, Crossref data citation.

Analyst-suggested research angles include institution/identifier matching, separating self-reuse from external reuse, field and publication-age comparisons, and manually validated classifications of sharing versus reuse statements. Any LLM assessment needs a labelled evaluation set, error analysis and reproducible model/prompt records. None of these specific design choices is advertised as already agreed. A validated corpus, an auditable classification workflow and a mixed-method account of credit are plausible outputs; a complete global census or proof that sharing causes economic benefit is not promised.

School, Programme and Field Context

The advertised host is the University of Edinburgh, School of Biological Sciences, with the BioRDM team in the C.H. Waddington building at the Centre for Engineering Biology. The official BioRDM facility supports research-data workflows and is supervised by Millar; the Centre describes a community of more than 50 groups and 200 researchers. These are relevant sources of research practice and collaboration, not a guarantee of unrestricted access to all institutional records or facilities. BioRDM, Centre.

The School reports fifth in the UK and first in Scotland for research power in THE's analysis of REF 2021, with 94% of research activity in the top two quality categories and a world-leading research environment assessment. Research power combines quality and breadth; this is a REF 2021-based biological-sciences measure, not a 2026 global ranking or a metaresearch-specific ranking. It supports a substantial biology research environment but cannot establish the quality of this individual PhD. School REF results.

The specified registration is PhD in Biological Sciences with Internship, starting 1 October 2027. EastBio is a four-year training programme. Core training is concentrated in years 1–2; years 3–4 put greater emphasis on research, alongside available professional/cohort activities. Its programme spans biological, mathematical and physical-science approaches. The project confirms mandatory training that includes a three-month placement; the current training calendar applies to existing cohorts, while the incoming cohort's detailed schedule and placement host require confirmation. Training structure.

Gary Polhill is based at the James Hutton Institute, a member of the EastBio partnership. Co-supervision does not create a joint degree, second registration or employment contract there. Edinburgh is the intended admitting university; exact registration regulations, progress reviews, thesis requirements and time spent at Hutton remain offer-specific questions. Partnership.

Supervisors, Recent Papers and Research Conditions

Millar's relevant contribution here is demonstrated by BioRDM's practical work and recent publications on data sharing and curation. The facility combines biologists with software developers and aims to integrate FAIR data management throughout research workflows. Polhill's official profile describes agent-based modelling of socio-environmental systems, rigorous model development and interpretation, policy scenarios and high-performance computing. This combination joins institutional data practice with computational and social-methods experience. It does not imply that the PhD must build an agent-based model. Polhill's contact is gary.polhill@hutton.ac.uk; Millar's recent paper gives andrew.millar@ed.ac.uk. The available profile and papers do not specify supervision frequency, current student load, completion outcomes or the division of day-to-day support. Polhill profile.

Three recent contributions

Deeb et al. (2025), “The rise of open data practices among bioscientists at the University of Edinburgh,” PLOS ONE, includes Millar. The study examines 555 papers from 2014–2023, distinguishing sharing completeness, accessibility, reusability and licensing. Some or all data were shared in 86% of papers; this does not mean complete sharing or downstream reuse. Its comparison with ODDPub also exposes limits of automated classification: after definition alignment, open-data sensitivity was 52% and specificity 90%; open-code figures were 30% and 97%. For this PhD, the paper provides institutional baseline evidence and a strong reason to validate LLM methods against manual assessment. Its sampled fields and scoring design limit generalisation. Paper.

Scorza et al. (2024), “Daily life in the Open Biologist's second job, as a Data Curator,” Wellcome Open Research 9:523, includes Millar. It describes curation alongside experimental researchers, organising biological datasets, clarifying metadata and licensing, and integrating manageable practices into daily work. Its contribution is operational evidence for how sharing quality is produced. It informs the institutional/ELN side of the PhD, but is not evidence that those outputs were later reused. Version 1 (12 September 2024) lists two approved reviews and one approved with reservations; it addresses open data rather than datasets with sharing restrictions. Institution-hosted paper.

Jager et al. (2025), “HUMAT: An Integrated Framework for Modelling Individual Motivations, Social Exchange and Network Dynamics,” JASSS 28(1):4, includes Polhill. It connects individual motives, cognitive dissonance, social decision processes and network influence, illustrating its framework through three applications. Its relevance is methodological: behaviour and adoption need interpretation, not only counts of visible events. It does not establish the causes of researcher data reuse, and HUMAT is not prescribed in the advert. Paper.

An older foundation, Ayllón et al. (2021), “Keeping modelling notebooks with TRACE,” Environmental Modelling & Software 136:104932, includes Polhill and proposes documenting iterative model development with structured notebooks. It explains the value of provenance when judging shared models. It is older than five years at this report date and is additional context, not one of the three recent papers above. Paper.

The public record demonstrates relevant methods and a functioning data-support environment. The project still needs explicit commitments about repository/ELN access, participant recruitment, ethics, computing and staff support. Access to internal research outputs cannot be assumed from an institutional affiliation.

Field Outlook and Career Implications

The project sits at a useful transition from monitoring openness to testing actual reuse. A 2026 author-deposited software record from Cohen and colleagues demonstrates an institutional citation-matching approach and explicitly interprets references as an indicator. It is useful feasibility evidence, not a complete definition of reuse or a supervisor paper. The cited April 2026 software record is distinct from a September 2026 release; the content of the project description's separate preprint DOI is unassessed, so no findings from that preprint are asserted. Archived approach.

The PLOS-commissioned March 2026 economic assessment argues that reuse-enabled infrastructure, standards, incentives and maintenance can produce efficiency and innovation benefits. It supports the importance of the question; macroeconomic claims cannot be transferred into a benefit estimate for this doctoral project. Economic assessment.

Career routes could include metaresearch, research-data stewardship/curation, scholarly infrastructure, research software and evidence/policy analysis. The evidence is stronger for relevant skills than for employment outcomes: UKRI's 2024 investment supports research technical professionals in software and data management, including data-professional communities and STEP-UP. It is not a vacancy forecast, salary benchmark or placement guarantee. A thesis with reproducible data pipelines, measurement validation and strong domain interpretation would develop portable skills; a literature-only account with no validated measurement would offer less distinctive technical evidence. UKRI workforce investment.

Funding, Application and Living Cost

Competitive awards and Chinese applicant eligibility

The four-year competition considers either an EastBio BBSRC Doctoral Landscape Award or the University of Edinburgh Doctoral College Award. The saved School guidance describes Doctoral College tuition coverage and a four-year stipend, with project research funds of £5,000 annually in years 1–3 and £1,000 in year 4. Research funds are separate from the living stipend; these are alternative award routes, not two personal incomes. Academic admission does not establish that an award is secured.

Successful applicants join the cohort as EastBio-funded or associate-funded students, with mandatory training including a three-month placement. The worldwide competition provides a UKRI-standard monthly stipend, tuition cover and project-cost support; Doctoral College Award consideration needs no separate application. Confirm the individual award and final 2027/28 rate. EastBio's fee-waiver/cap and study rules below do not automatically establish identical conditions for an associated Doctoral College award or guarantee a particular placement host.

EastBio accepts international applicants and sets a 30% annual cap on international intake across the network; this is not this project's quota or an individual applicant's probability of winning. Successful international EastBio-funded awardees receive a waiver of the overseas tuition-fee difference. Whether the separate Edinburgh Doctoral College Award route covers this difference remains unconfirmed. Visa fees and the immigration health surcharge are excluded. Visa-requiring students must register full-time and live within commuting distance except during permitted research or placement periods. Confirm the actual offer's funding route, full award-period tuition coverage and personal relocation costs. Programme FAQ.

The 2026/27 UKRI minimum stipend is £21,805 annually, about £1,817 monthly, and the home-fee contribution is £5,238. These are benchmarks, not the confirmed 2027/28 stipend. UKRI rates.

Application steps and conflicting instructions

The supplied project route is EUCLID, selecting Biological Sciences with Internship and 1 October 2027. Include the project title, proposed supervisors and anticipated funding source. Upload the completed EastBio application form into the Research Proposal section; this placement of the form is not an instruction to author an additional proposal. Give referees the official EastBio form in advance. EUCLID can email an upload link when referee details are entered, but the letter must still use that template.

The project says midday Monday 16 November 2026, allows reference letters for two weeks afterwards, and keeps applications under consideration until 22 February 2027, when shortlisted candidates are invited to interview. Missing information may be requested through EUCLID; monitor both portal and email. The university reserves discretion to amend or withdraw advertised scholarships without further notice. The exact noon time zone and the reference allowance need host confirmation because the central programme asks for references by 16 November.

The School's general funded-PhD guidance allows up to three projects and describes eligibility-checklist/EUCLID steps. EastBio's specific rule allows only one EastBio project, which controls this competition. Central instructions also require the anonymous EDI survey response ID on the form, transcripts and two official-form references; project supervisors cannot be referees. A 2:1 or equivalent is the standard programme qualification, with alternative consideration possible for a master's or significant relevant experience. The seed advert did not state a specific subject/grade or coding requirement, so confirm international equivalence and School English-language conditions rather than invent them. School admissions.

The published EastBio selection process includes a short presentation about previous research and questions about the project, research potential, interdisciplinary awareness and programme commitments. The central timeline lists 22 February invitations and March interviews. A local invitation controls final arrangements. Selection process.

Edinburgh costs and stipend coverage

The University's official living-cost page provides an indexed 2026/27 budget of £18,504 per year, or £1,542 per month, for a single postgraduate, excluding tuition. Monthly categories are accommodation/utilities £978, food and other shopping £248, mobile £12, social/leisure/wellbeing £246, and local buses for age 22+ £58. Housing type materially changes costs. The figures were calculated in March 2026 and are planning guidance, not a 2027 quote; they come from the official page's indexed text, so confirm the current page and obtain a personal housing quote. Official living-cost budget.

Against the above 2026/27 £21,805 stipend benchmark, the annual difference is £3,301, about £275 per month. This is a relatively modest margin before relocation, visa/IHS, emergencies or costs omitted from the illustrative budget. It is not guaranteed saving, especially for an October 2027 start, different accommodation or dependants. Obtain a real full-year housing quote and the offer-year stipend; do not equate a visa maintenance threshold with a sufficient Edinburgh budget.

Applicant Fit and Preparation

The work suits an interest in computational analysis and research behaviour. Data science, bioinformatics, environmental informatics, quantitative social science and bioscience research experience may provide different entry strengths; these are analytical background mappings, not an official closed degree list. The advert does not specify programming languages, prior LLM publications or interview-method expertise as hard requirements. No distinct preferred-skills list was verified.

Useful preparation includes a reproducible analysis example, identifier/metadata cleaning, basic text classification and statistical evaluation, and a clear account of why a citation may not equal reuse. Python/R, APIs, version control and annotation practice are suggestions. If qualitative work is pursued, participant consent, ethics and careful integration with quantitative evidence will matter. Familiarity with biology's heterogeneous data helps interpretation; this is not advertised as a conventional wet-lab doctorate. Applicants seeking primarily to invent new foundation models, or exclusively bench experiments, should clarify the scope before choosing this project.

Next Actions and Questions to Verify

Before applying: discuss the four research questions with Millar and identify which outputs/fields are prioritised. Ask the School to confirm the exact project is open, noon time zone, programme selection, funding consideration, EDI step and whether the two-week reference extension applies. Arrange both references for 16 November if feasible while awaiting clarification. Ask how international academic and English-language requirements apply.

At interview: establish what repositories and ELN material can actually be accessed, permissions and contingency data; who defines and manually validates reuse; whether code/models require different measures from datasets; how LLM error and field bias will be evaluated; and whether surveys/interviews are planned or optional. Clarify ethics approval, participant recruitment, computational budget, RSpace collaboration terms and the division of Millar/Polhill and BioRDM staff support.

Before accepting: secure written confirmation of award route, 2027/28 stipend, full tuition arrangements, research/placement costs, registration milestones and any Hutton travel. Check publication/data-access rights and what happens if internal access or recruitment is delayed. Compare the offer with actual housing costs and visa/IHS reserves. A good opportunity here depends on a credible, accessible measurement programme as much as institutional reputation.

Evidence and Sources

The School funding page lists a general 19 October–16 November 2026 competition; that window does not establish the present status of this specific project.

Project-specific tasks and local application instructions are provided in the project record; the linked EastBio rules give programme-level guidance. School funding, EastBio application rules.

Host, school and training

• School funded-PhD admissions
• School funding and 2026 competition
• BioRDM facility
• Centre for Engineering Biology
• School REF 2021 results
• EastBio partnership
• EastBio training

Supervisors and research evidence

• Gary Polhill profile
• Deeb et al., 2025: institutional data-sharing practices
• Scorza et al., 2024: data-curation practice
• Jager et al., 2025: HUMAT framework
• Ayllón et al., 2021: TRACE modelling notebooks
• Data Citation Corpus documentation
• Crossref data-citation practice
• Cohen et al., 2026: archived institutional reuse approach
• 2026 economic assessment of open science

Eligibility, finance and careers

• EastBio eligibility and funding FAQ
• EastBio application instructions
• EastBio selection
• UKRI stipend benchmarks
• Edinburgh postgraduate living-cost budget
• UKRI digital/technical workforce investment

Ordered Bibliographies and Version Limits

1. Haya Deeb; Suzanna Creasey; Diego Lucini de Ugarte; George Strevens; Trisha Usman; Hwee Yun Wong; Megan A. M. Kutzer; Emma Wilson; Tomasz Zielinski; Andrew J. Millar. The rise of open data practices among bioscientists at the University of Edinburgh. PLOS ONE 20(7):e0328065, 23 July 2025. Millar is tenth of 10 authors; the measurement findings and limitations are discussed above. Paper.

2. Livia C. T. Scorza; Tomasz Zielinski; Irina Kalita; Alessia Lepore; Meriem El Karoui; Andrew J. Millar. Daily life in the Open Biologist’s second job, as a Data Curator. Wellcome Open Research 9:523, version 1, 12 September 2024. Millar is sixth of 6; this version has two approved reviews and one approved with reservations, not an inferred later-version status. Paper.

3. Wander Jager; Patrycja Antosz; Loes Bouman; Teng Li; J. Gareth Polhill; Timo Szczepanska; Shaoni Wang. HUMAT: An Integrated Framework for Modelling Individual Motivations, Social Exchange and Network Dynamics. JASSS 28(1):4, 31 January 2025. Polhill is fifth of 7; contribution and correspondence roles are not inferred from byline order. Paper.

4. Keeping modelling notebooks with TRACE. Environmental Modelling & Software 136:104932, February 2021. The saved author list is incomplete; known authors include Daniel Ayllon, Steven F. Railsback and J. Gareth Polhill. This older methodological background is not a fourth recent contribution. Paper.

Supplementary Recruitment References

• Deeb study DOI
• Cohen preprint DOI: content not assessed
• TRACE DOI
• Andrew Millar university profile