Uncertainty-aware structure-to-specificity modelling for generalizing transcription factor binding predictions from sparse mutational data

Recent structure-based computational approaches have shown promise for predicting protein–DNA binding specificity. DeepPBS, a geometric deep-learning model, predicts binding specificity directly from protein–DNA structures, including AlphaFold-predicted ones and has been applied to designed proteins targeting specific DNA sequences (Mitra et al., 2024). Separately, FamilyCode uses a reference-free representation of binding-preference variation within a structural family to predict how mutations shift TF binding free energy, demonstrated on the bHLH and homeodomain families (Liu et al., 2025, Nucleic Acids Research, doi:10.1093/nar/gkaf831). Neither approach has been applied to bacterial transcription factors, which are at the heart of engineering biology. Further, neither approach has been benchmarked against a dataset with this project's depth of per-variant characterization: we have experimental data on 30 structural TetR variants, each measured for binding across 100,000s of DNA sequences.

This project asks whether a model can be built that maps AlphaFold-predicted structural perturbations in TetR variants to shifts in DNA-binding specificity, and reliably quantifies its own uncertainty when extrapolating to synthetic variants with varying sequence distance from the 30 measured ones. The central challenge is one of generalization from a few labelled examples: while each of the 30 variants is characterized against a rich binding landscape, the axis the model must generalize across (i.e., how structural change maps to specificity change) has only 30 data points. In other words, the project is not about a data-scarcity problem but about a small-N transfer learning problem sitting on top of large, well-resolved per-variant measurements.

The project will develop machine learning methods across three stages. First, structure-effect modelling: characterizing how point mutations perturb TetR structure, addressing AlphaFold's known insensitivity to single substitutions, likely requiring single-sequence rather than multiple sequence alignment inference, and protein-language-model-based mutation scoring rather than standard AlphaFold2/3 prediction. Second, per-variant specificity modelling: fitting biophysical/thermodynamic models to each of the 30 variants' individual binding datasets, establishing a high-resolution specificity landscape per structural perturbation. Third, cross-variant generalization: learning a structure-to-specificity mapping from the 30 labelled examples, benchmarked directly against previously published approaches (DeepPBS, FamilyCode). Lastly, uncertainty quantification: producing calibrated confidence estimates for each synthetic variant's predicted binding profile, feeding an active-learning loop that proposes which variants are most informative to experimentally test (to be carried out by a senior technician in Lagator lab).

Put together, the project aims to close the loop between structural prediction and experimental design. As a result it will transform TetR engineering from empirical trial-and-error into a targeted, uncertainty-guided search.

Desirable Student Background:

Key: computer science/maths/statistics. Desirable: experience with protein structure prediction (AlphaFold or similar), Python-based ML/scientific computing, and ideally some exposure to Bayesian methods or uncertainty quantification. While molecular biology background to interpret transcription factor–DNA binding data is useful, the supervisory team and our research groups more broadly provide plenty of expertise in those areas. Hence, a computational background with strong interest in learning biology is ideal.

How to apply:

Please apply under the University of Manchester application portal and select 'PhD in Artificial Intelligence' (OAA Application portal)

When applying, you’ll need to specify the full name of this project, the name of your supervisor, if you already having funding or if you wish to be considered for available funding through the university, details of your previous study, and names and contact details of two referees.

Your application will not be processed without all of the required documents submitted at the time of application, and we cannot accept responsibility for late or missed deadlines. Incomplete applications will not be considered.

You will be asked to upload the following supporting documents:

• Final Transcript and certificates of all awarded university level qualifications
• Interim Transcript of any university level qualifications in progress
• CV
• Supporting statement: A one or two page statement outlining your motivation to pursue postgraduate research and why you want to undertake postgraduate research at Manchester, any relevant research or work experience, the key findings of your previous research experience, and techniques and skills you’ve developed. (This is mandatory for all applicants and the application will be put on hold without it).
• Contact details for two referees (Please ensure referee email addresses are official university or workplace accounts, as we may need to verify the references.)
• English Language Certificate (if applicable)

We actively encourage applicants from diverse career paths and backgrounds and from all sections of the community, regardless of age, disability, ethnicity, gender, gender expression, sexual orientation and transgender status.

We also support applications from those returning from a career break or other roles. We are dedicated to supporting work-life balance and offer flexible working arrangements to accommodate individual needs. Our selection process is free from bias, and we are committed to ensuring fair and equal opportunities for all applicants.

We are dedicated to supporting work-life balance and offer flexible working arrangements to accommodate individual needs, including a part-time study option.

Join us and help grow the UK’s national capability in artificial intelligence, part of the UK’s modern industrial strategy. Everyone is welcome if you are passionate about uncertainty quantification, decision making with humans in the loop and decision making for machine learning systems and their applications in Science and Engineering, and looking to develop your research and innovation skills at the doctoral level to further your career in the AI sector.

Equality, diversity and inclusion is fundamental to the success of The University of Manchester, and is at the heart of all of our activities. We know that diversity strengthens our research community, leading to enhanced research creativity, productivity and quality, and societal and economic impact. (Equality, diversity and inclusion | The University of Manchester)

Funding Notes

Fully funded AI UKRI CDT 4 year program; The Project will be based in University of Manchester.

The UKRI AI CDT in Decision Making for Complex Systems is part of the TechFirst offering stipends of £30,000 for students eligible for home fee status. The start date is September 2027.

References

Mitra, R., et al. Geometric deep learning of protein–DNA binding specificity. Nat Methods 21, 1674–1683 (2024). https://doi.org/10.1038/s41592-024-02372-w
Liu, S., et al. Predicting the DNA binding specificity of transcription factor mutants using family-level biophysically interpretable machine learning, Nucleic Acids Research, Volume 53, Issue 16, gkaf831, https://doi.org/10.1093/nar/gkaf831
https://research.manchester.ac.uk/en/persons/mato.lagator/
https://research.manchester.ac.uk/en/persons/magnus.rattray/
https://research.manchester.ac.uk/en/persons/simon.lovell