Machine learning transition state geometries

   Centre for Accountable, Responsible and Transparent AI

This project is no longer listed on and may not be available.

Click here to search for PhD studentship opportunities
  Dr Matthew Grayson  No more applications being accepted  Self-Funded PhD Students Only

About the Project

We have an exciting opportunity to work on the development of new transition state prediction methods. You will join the Grayson group in the Department of Chemistry to develop fast and accurate transition state geometry prediction tools by combining machine learning (ML) with reaction modelling for applications in pharmaceutical chemistry.

As a PhD student, you will:

·        Gain experience in the use of ML and reaction modelling for transition state geometry prediction.

·        Perform both individual and collaborative research projects.

·        Write up research results for publication in scientific journals.

·        Present your work at national and international conferences.

Project Details:

An unmet need in computational chemistry is the development of reaction modelling methodology that is fast and accurate and provides mechanistic insight from transition state (TS) geometries; existing modelling and machine learning (ML) methods compromise on at least one of these criteria. Such an approach to predicting reactivity would provide an alternative to costly experimental screening in synthetic route design and optimisation.

TS geometries are of particular importance in reaction design given their role in rationalising reaction outcomes, mechanistic elucidation, and improving the accuracy of ML reaction barrier prediction models in low data regimes. However, the standard approach to generating TSs with reaction modelling (e.g. quantum mechanical (QM) calculations) is highly time-consuming which prevents the screening of large numbers of reactions and conditions.

In this project, a novel approach to the rapid generation of QM-quality TS geometries using ML will be developed. These ML models will deliver high-quality TS geometries in seconds which is many orders of magnitude faster than QM. We will also build ML models that deliver fast and accurate reaction barrier predictions. Use of these TS and barrier prediction models will help eliminate experimental trial-and-error and provide a more rapid and sustainable approach to reaction discovery.

Further information about the Grayson group and our research interests can be found at and in references [1] and [2].

Applicants should hold, or expect to receive, a master's degree or first or upper-second bachelor's degree in a relevant subject. We are looking for a highly motivated individual to join our group. Experience with coding (any language) is desirable but not essential. Experience with machine learning and reaction modelling is not essential.

Formal applications should be made via the University of Bath’s online application form. Further information about the application process can be found here.

Start date: Between 8 January and 30 September 2024.

Chemistry (6) Computer Science (8) Mathematics (25)

Funding Notes

We welcome applications from candidates who can source their own funding. Tuition fees for the 2023/4 academic year are £4,700 (full-time) for Home students and £26,600 (full-time) for International students. For information about eligibility for Home fee status:


[1] Machine learning activation energies of chemical reactions. Wiley Interdiscip. Rev. Comput. Mol. Sci. 2022, 12, e1593.
[2] Machine learning and semi-empirical calculations: A synergistic approach to rapid, accurate, and mechanism-based reaction barrier prediction. Chem. Sci., 2022, 13, 7594.

How good is research at University of Bath in Chemistry?

Research output data provided by the Research Excellence Framework (REF)

Click here to see the results for all UK universities

Where will I study?

Search Suggestions
Search suggestions

Based on your current searches we recommend the following search filters.