Dr Nikolas Pontikos is a Senior Research Fellow at UCL Institute of Ophthalmology, and has been working on inherited retinal conditions for several years. He is a bioinformatician, harnessing computing power to analyse vast amounts of healthcare and scientific data, generating answers in far less time than would otherwise be possible.
“Growing up with computer games in the 90s and witnessing the information revolution brought by the internet, I started getting really interested in maths and computers in my late teens, which led me to an undergraduate degree in computer science” Dr Pontikos told us. “Sadly, towards the end of my degree, my mother received a late cancer diagnosis and passed away. This made me realise how important it was to improve healthcare, and that I was in a position to contribute to this area with my computing knowledge.”
Dr Pontikos therefore married up his computing skill with biology and medicine via postgraduate qualifications (bioinformatics MSci at Imperial College and PhD at Cambridge University) and work at the European Bioinformatics Institute, before arriving at UCL to work with Dr Vincent Plagnol on rare disease genetics. This led to an introduction to Prof Alison Hardcastle, who needed bioinformatics support for the UK Inherited Retinal Disease Consortium (UKIRDC) project, funded by Retina UK.
Dr Pontikos took charge of the UKIRDC’s bioinformatics pipeline, which processed the genetic data from everyone who had contributed a DNA sample to the project. “The computer systems originally refined by Dr Plagnol were able to identify different types of disease-causing genetic changes, and we used this to help us spot changes that had the potential to disrupt retinal function” Dr Pontikos explained. “Nearly everyone living with sight loss who took part in the UKIRDC had previously been screened for faults in genes already known to be associated with retinal disease, without success, and we were therefore often able to identify completely novel disease-causing genes. These discoveries can provide families with a clear genetic diagnosis and enable scientists to identify potential treatment pathways.”
Dr Pontikos and his team are now leading some large projects of their own, including an innovative endeavour called Eye2Gene, the idea for which grew directly from his work on the UKIRDC. The Eye2Gene project aims to develop an artificial intelligence system that can use retinal scans alone to identify the likely disease-causing gene, making genetic diagnosis more readily and rapidly available to everyone with an inherited retinal condition.
Dr Pontikos has been awarded a significant grant by the National Institute for Health Research (the research arm of the NHS) to develop Eye2Gene. “The Eye2Gene AI system needs to learn from the retinal scans of people with a known genetic diagnosis” he explained. “So far, it has only been trained on genes for which at least 10 people’s scans were available, covering a total of the 36 most common inherited retinal disease associated genes. In order to optimise its performance and extend it to more genes, we will need to increase the amount of data to at least 20 patients per gene. This requires us to gather data from several sites in the UK and internationally.
“Our grant runs until 2024, and by that time we will have developed the technology and tested it sufficiently to prove its performance and utility, so I would expect it to be used as a research tool in the next three years. It will take more extensive testing over an extra couple of years before it can be used day-to-day in the clinic.”
Dr Pontikos acknowledges that working with rare conditions is not without its challenges. “From a practical perspective, the datasets are relatively small, and in addition, rare disease does not necessarily attract the same level of research funding as other conditions. That’s why the support of charities like Retina UK is really important” he said.
However, there are plenty of positives too. “One of the most exciting moments was finding out that Eye2Gene performed very well on external datasets, showing that the algorithm is able to generalise” he told us. “In addition, the interest and support from the inherited retinal disease community, and the motivation of my team, is truly inspiring. We are very grateful to all those living with sight loss who have provided DNA samples and retinal scans”