Our UKRI Innovation Fellows and UK HDR Rutherford fellows received three years of funding to pursue their research goals and transform from being post-doctoral researchers to independent researchers.

All had bespoke training, mentorship, leadership opportunities, interdisciplinary training resources, plus the chance to take part in a programme of events and workshops.

We wanted to see how the fellowships have affected their lives and careers. The first we caught up with are Honghan Wu and Rona Strawbridge.


Honghan Wu – From Gamer to Professor

Honghan Wu has crossed many borders and disciplines to find an area of work that makes full use of his abilities and interests. A crucial stage in his career development was the HDR UK Rutherford Fellowship which was a stepping stone to towards his new professorship. We spoke to him about a career that has included everything from being a semi-professional gamer to helping transform the use of medical records in research.

Tell us about your life before health data research

My undergraduate studies were in China, in mechanical engineering. My university, Southeast University in Nanjing, was one of the first to use a lot of computer programing (coding) in undergraduate programmes, and I got super-interested in coding and computers in general. After graduation I never did anything mechanical engineering related. Instead I went to work for a Taiwan-based company, Inventec, which was developing a new branch in PC gaming.

We had to play a lot of games to understand what people like about them – you have to relate to the logic behind the games, there are mathematical calculations as well, just like in chess.

It turned out I was very good at one called StarCraft. I ended up as a semi-professional player.

Sounds an interesting career – what changed?

I had a good salary and was progressing fast. Like many big Asian companies it had a clear career progression ladder – I could see exactly where I would go over the next 10 or 20 years. But I found it a bit difficult to get my voice heard in such an enterprise and I had a lot of thoughts.

So, I didn’t find that very inspiring and thought I could do something more exciting – maybe go back to university, learn about really big ideas, do some research, change the world somehow.

So what did you do?

I took a master’s, then a PhD in Southeast University, Nanjing whch was focused on the so-called semantic web, and on ways of using AI to make computers smarter. Then I came to the UK in 2013, to Aberdeen University, as a Post Doc on a European funded project which had an exemplar use of semantic web technologies in healthcare. This was a collaboration with IBM Italy.

They had lots of data from an oncology hospital in Milan. The patient records showed that people’s treatment often varied from the guidelines. We were trying to figure out why. It could be that a patient is too ill for surgery or some other reason.

But it is really important in creating guidelines that you account for all likely scenarios that clinicians will face. That means you have to be able to accurately capture real-world information.

After that I moved to King’s College, London where I was one of the first developers working on CogStack. We started with access to just 500 patient records and were looking at fundamental issues. Does this technology makes sense? Would it be safe to apply to all records?

Around that time the government was looking to recruit 100,000 patients (and their relatives) for a study on rare diseases and cancer. Normally that would involve getting a research nurse to read lots of patient referall letters to find suitable people – a very slow process.

We used AI to go through the letters and identify key words and sentences, which made it much quicker. That was a big success for our team as it was a real-world project.

How did your HDR UK Fellowship come about?

It was about this time that HDR UK was offering fellowships and I applied for the one at Edinburgh. After the successes we had with CogStack at King’s College London it was a great opportunity to test whether the technology was applicable at a different site and in a different health system.

Was your fellowship looking at health records?

Yes, it was about making the most out of records, bringing together disparate information to get the fullest picture for better care.

The problem is that healthcare information is scattered and in different forms, it’s not all linked. And some formats are much easier for computers than others. Lab test results on spread sheets are straightforward, doctors’ letters are not.

First, the computer has to understand people’s language and that’s not easy. Even when discussing the same symptoms, different words can be used. And there are differences between what doctors might say in England or Scotland, and even slightly different jargons between hospitals.

And contextual information is at the heart of computer capture. If the computer is looking for records of cancer patients it needs to distinguish between the record saying a patient has cancer and one that says a patient is worried that they have cancer.

My work involved trying to interpret the many different types of data and bring them together.

What did fellowship lead to?

First, I got got a UCL lectureship – a natural transition from personal fellowship to academic position. And recently I was appointed as Glasgow University as a Professor of Health Informatics and AI at the School of Health and Wellbeing.

I have always had the ambition to be able to work on data at a national level. Scotland is great for that because it has a vast amount of linked health data with great quality.

Can you give an example of what you hope to achieve at Glasgow?

One example is imaging data. In Scotland it has all been gathered in one place for the last 15 to 20 years. Millions of records that are underused. It’s a massive resource of real-world information about diseases, diagnoses and treatments that would be enormously valuable for research and care.

But there are many challenges – there is a great deal of free text, such as radiologists’ reports or doctors’ notes that have to be dealt with. Also, there is a great deal of sensitive information which could lead to patients being identified. It all has to be very well anonymised.

What we are looking at doing is paraphrasing everything so researchers have the data they need but all the sensitive material is removed.

Would you encourage people from other other backgrounds to consider health data science?

Absolutely. It’s an area where you can apply your knowledge and skills and see the impact on people’s health, you can save lives. And there are so many areas of work. Things like robotics, AI and virtual reality are developing very fast, so this is a great time to get involved.

And we need people with different knowledge and backgrounds, like engineers – people who can make devices smaller and easier to use.

And if you are interested in the business potential it’s also a great area. For example AI in medicine might well be the next big thing, and that’s going to mean lots of exciting new business start ups in years to come.

Do you still enjoy gaming?

Yeah, I still love StarCraft actually and am following the e-sports on that and alike.


Rona Strawbridge – Advancing Research into Mental Illness

One of the great benefits Dr Rona Strawbridge gained from her fellowship was the chance to carry out independent research. As a senior lecturer at the University of Glasgow, this continued freedom is something she values enormously – allowing her to pursue her own ideas on the links between mental illness and cardiometabolic diseases.

How did you come into health data research?

I did a BSc in biochemistry with medical biochemistry, which gave me a broad education into how health and disease can be studied and understood.

I particularly enjoyed learning about genetics, the logic of which suits my brain. So, picking a dissertation project, then a PhD project using genetics was the obvious next step.

The data part (as we understand it today) was not really considered then, as most of the data that would be studied would also be generated (in a manageable quantity) in a lab by researchers.

This changed during my PhD, when the first genome-wide association study was conducted. By the time I started my first post-doc, genotyping costs had reduced enough that there was an explosion in this type of data, and projects using the data.

Once the number of variants being analysed was greater than Excel can handle (ie one million!), and it became clear that there would be exponential growth in this area, it also became clear that data science methods would be needed. So, for me, the data part came long after the health research part.

What led to the fellowship?

My post-doc positions at Karolinska Institute focused on using genetics to understand obesity, type 2 diabetes and cardiovascular disease and their inter-relatedness. When I moved to the University of Glasgow, I switched to using genetics to investigate mental illness. The fellowship was an opportunity to combine both areas of interest.

What did you do for your Fellowship?

I was able to apply the same logic that has been used in cardiometabolic diseases to mental illnesses, that is to consider the features rather than just the diagnoses. So, in the same way that glucose levels have taught us about diabetes, then investigating risk-taking and suicidal behaviour can tell us about mental illness.

Some of my work has been identification of genetic regulators of risk-taking and suicidal behaviour. Some of my work has systematically looked at whether the same genetic regions influence mental illness and cardiometabolic diseases, which is hard to fully assess from publicly available data.

I have also repurposed a standard quality control method, to identify subgroups of individuals with different phenotypic presentation. More recently, I led an international collaboration that demonstrated that increased genetic risk for depression is associated with increased risk for blood clots.

Did it open up new or enhanced career opportunities?

Independence to follow my own research interests was definitely the biggest opportunity for me. Enhanced career opportunities are less obvious, but the training/resources/travel budget allowed me to upskill in data science methods, access to a number of useful datasets and attend conferences that lead to collaborations. Meeting other fellows has been another fantastic benefit, providing a hugely valuable peer support network.

What you are doing now?

I am now a senior lecturer at the University of Glasgow, which means I have a teaching commitment and I supervise MSc projects. Still, I have the independence to conduct my own research and have PhD students working with me on the link between mental illness and cardiometabolic diseases.

I am applying for new funding opportunities, particularly to extend the sub-grouping work and increase the diversity of data I am using. I co-lead the Behavioural Epidemiology and Genetics research group, which is more a research “collective” than a traditional group.

Where would you like your career to go in future?

In some ways I would like to continue just as I am. On the other hand, a more senior role does involve more responsibilities that take time away from research.

I have too many ideas, but not enough time to work on them. So ultimately getting more funding would allow me to recruit others who are equally enthusiastic about understanding the link between mental illness and cardiometabolic diseases, so that more research can be carried out.

Another slightly different ambition is to promote a positive and supportive research environment. I am lucky to be in such an environment and want to feed that forward as much as possible.


Our Fellowships

You can also find out more about our UKRI Innovation Fellows and UK HDR Rutherford fellows here. While this programme has now closed we have since launched the HDR UK fellowships linked to our Research Driver Programmes.