Accelerating Open-Source Chemistry Tools Through IAA Funding
When Dr Fernanda Duarte joined the Department of Chemistry at the University of Oxford in 2018 as an Associate Professor in Computational Organic Chemistry, she set out to build a research programme at the intersection of theory, computation, and real-world chemical problems.
Leading a team of PhD students and postdoctoral researchers and teaching as a Fellow of Hertford College, her work sits at a key interface: translating sophisticated computational methods into practical tools that chemists can actually use.
A small but strategic pot of Impact Acceleration Account (IAA) funding became a turning point, not only for her research group and its collaborations with industry, but also for the career of her PhD student, Tom Young, and for a growing community of users in both academia and pharma.
Fernanda DuarteFrom PhD Prototype to Production-Quality Tool
The IAA funding was awarded to support the further development of a piece of software, named autodE, that Tom Young had originally created during his PhD in Duarte’s group.
industry, particularly with AstraZeneca. The original plan included visits and close collaboration; however, the COVID‑19 pandemic disrupted many of those in-person elements.
The core idea was simple but powerful: chemists frequently need to understand how a chemical reaction proceeds from reactant A to product B, including the energetic pathway and the transition states along the way. Traditionally, this involves a great deal of manual, expert-driven computational work.
The project aimed to change that, as Duarte explains: “The goal was to develop a piece of software to automate how we model chemical reactions, requiring researchers to input only a few lines to run their calculations. This produces an energy profile to quantitatively analyse how one goes from A to B”
While the IAA funding did not radically transform the scientific basis of the original PhD software, instead it did something equally important for long-term impact: it allowed the team to refactor, clean, and harden the software into a robust, high-quality, user-friendly tool. Crucially, the funding bought dedicated time to move from “it works for the person who wrote it” to “it works for many people who didn’t write it.” This shift has been central to the software’s continued use and adoption.
Building Bridges to Industry: AstraZeneca and Beyond
From the outset, the project was conceived as a mechanism to deepen engagement with industry, particularly with AstraZeneca. The original plan included visits and close collaboration; however, the COVID‑19 pandemic disrupted many of those in-person elements.
Despite the disruption, the IAA support still catalysed tangible engagement with AstraZeneca, and in addition, the work directly enabled industry-linked industrial Case (iCase) studentships with both AstraZeneca and IBM, while the software has attracted interest from other major firms, including Syngenta.
Duarte describes how the tool has become part of a broader ecosystem of collaboration: “It was through the publication of this software that AstraZeneca researchers approached us and invited us to apply for an iCase, which was successful and 100% funded by AstraZeneca. Now, this student has continued developing this code; he has recently visited AstraZeneca, working within the R&D team to install and use the software. So the value and legacy of the IAA funding continues to be applied in a very practical way.”
Beyond AstraZeneca, she notes: “Other agricultural technology firms such as Syngenta have also contacted us, and they are now using the software in some of their research projects too.”
Although there has not yet been large-scale follow-on funding purely focused on the software itself, the tool has become an enabling platform for other projects and partnerships, embedding itself in the group’s ongoing research and external collaborations.
Engaging with RisingWISE
Fernanda’s own journey through entrepreneurial learning has also shaped her approach and understanding of innovation in research. As she notes of her participation in the EPSRC IAA-funded RisingWISE programme for Oxbridge women in STEM:
“RisingWISE was a formative experience for me. It helped me learn how to present my science in clear, simple terms, and made me realise how important it is to ‘pitch’ ideas, not just do the technical work. Being in a women-only, science-focused group where people really understood the pressures of academia showed me the power of dedicated support networks. It also opened my eyes to skills like negotiation and self‑advocacy, which I now actively encourage in my students and peers, both in academia and in collaboration with industry.”
Open Source, with a Thoughtful Approach to Value
From the beginning, Duarte and her team were keen that the software should be an open-source tool, accessible to researchers who might not have the resources to purchase commercial solutions.
Their tool was designed not to lock users into a single proprietary ecosystem. Instead, a key design principle was interoperability. The intention was that it should interface with other existing software so there were fewer constraints on its applications and uses.
Over time, however, success has brought new questions. As more companies began to show interest and potentially build on the software for commercial purposes, the team realised they needed a more nuanced licensing and distribution model.
While they have not yet fully explored commercialisation, the groundwork has been laid: an openly accessible academic tool with potential pathways to revenue-generating use in industry.
Transformative Impact on a Student’s Career
One of the most striking outcomes of the IAA-funded project is its impact on Tom Young, the PhD student whose software formed the basis of the work.
Through the project, Tom began collaborating closely with the Research Software Engineering (RSE) team at Oxford. This experience not only improved the software but also directly influenced his career trajectory. Through this, he decided to pursue a career in software development, and today he is one of the lead people in the equivalent team at UCL.
Duarte notes that Tom found out about the role via the RSE team in Oxford, as she says, “the RSE community in the UK is growing and supportive of each other.”
Even after graduation, Tom remains actively involved in his free time. He continues to support the group with the software, and is an informal mentor and technical reference point for current students.
Duarte also notes that Tom’s standards still shape the team’s approach to software quality, and although he is humble and not eager for formal recognition, she feels that he exemplifies how impact funding can launch careers that blend scientific and software expertise.
Personal and Institutional Learning: Discovering Hidden Resources
For Duarte herself, the EPSRC IAA project was more than a technical upgrade; it was an introduction to a new way of working within the University.
When she first arrived in Oxford, she was not fully aware of the support available for research software: “At the time that we were working on this, around five years ago, it felt like the coding aspect was not highly valued, and we had to do everything ourselves.”
Preparing the IAA application changed that. She discovered the Research Software Engineering team, realised how critical their expertise could be, and built them into the project: “With this funding, we were able to work with the engineers in Oxford… and that was very important… I knew, while preparing the application, that this would be a critical aspect to include.”
This collaboration had several lasting effects. It lifted the quality and longevity of the software, and it has raised the profile of Duarte’s group within the department and the wider University. Perhaps most importantly and more far-reaching, it has highlighted the value of long-term planning in infrastructure-style research outputs, rather than one-off prototypes.
More broadly, the process exposed Duarte to the hidden networks and expertise within the University: “At Oxford and other institutions, many wonderful things are often happening in silos that you don’t necessarily get to hear about. Unless you have that interdisciplinary opportunity, it can be difficult to know what’s out there and how you can collaborate.”
Building on the Foundation: Next-Generation Tools with Machine Learning
Today, the original software, autodE, is no longer a standalone project; it has become a core infrastructure component for Duarte’s broader research agenda.
“Today we continue developing autodE, using it as a core part for other software, mlp-train and cgbind…it is very much a key part of the work that we do.”
Dr Fernanda Duarte
The next frontier is to scale from idealised reactions to more complex systems in solution, incorporating machine learning interatomic potentials (MLIPs) to approximate solutions to the Schrödinger equation more efficiently:
Although Duarte does not describe herself as a computer scientist: “Machine learning today is a key part of our research, but this is not something I had originally planned to take up. I’m not a computer scientist, but I am very invested in exploring how we can use the emerging AI technologies to solve long-standing chemistry problems.”
Conclusion
The EPSRC IAA-funded software has made it possible to plug new methodologies into an existing, trusted framework, accelerating the path from method development to usable tools.
What began as a small, targeted investment in “just a piece of software” has evolved into a sustainable, evolving foundation for computational chemistry innovation, demonstrating how well-placed impact funding can amplify both scientific and human potential.