Michael Skinnider wins the ICTP-IBM Brahmagupta Prize for Artificial Intelligencefederica.dauria Ven, 07/24/2026 - 10:18 28 Luglio 2026 Language Italian Federica DʹAuria “For pioneering the application of artificial intelligence to biomedical discovery, and for developing transformative AI methods that reveal previously unknown chemistry, accelerate metabolite discovery, advance neuroscience, and address major challenges in human health”, Michael Skinnider, a computational biologist at Princeton University, has been awarded the ICTP-IBM Brahmagupta Prize for Artificial Intelligence. The prize was presented on Thursday, July 23rd, during a ceremony at the Abdus Salam International Centre for Theoretical Physics (ICTP) in Trieste. The prize – named after the Indian mathematician and astronomer Brahmagupta – was established in 2024 to recognise young researchers who have achieved outstanding results with promising benefits for society in the field of Artificial Intelligence. Skinnider came to computational biology after studying medicine, and is now a professor at Princeton University, where he applies machine learning methods to the study of biochemical processes. In particular, he uses artificial intelligence to study special molecules found in our bodies – known as metabolites – and the ways they influence human health. The lab he directs at Princeton's Lewis-Sigler Institute aims to shed light on the so-called “dark matter” of the metabolome (the full set of metabolites present in the human body) which current techniques can detect but not yet identify in detail. SCIENZA E RICERCA © Roberto Barnabà / ICTP Photo Archive We spoke with Skinnider at the ICTP in Trieste on the occasion of the award ceremony, and asked him to walk us through the main aspects of his work on metabolites, and more. What are metabolites, and why, as you argued in a study published in Nature earlier this year, is there still so much we don't know about them? “Most readers will probably be familiar with DNA, RNA and proteins, which are polymers made up of nucleotides or amino acids. Metabolites are much smaller molecules (the term “metabolite” is in fact often used interchangeably with 'small molecule') and they include substances such as glucose, amino acids and vitamins. They are, in other words, the “building blocks” for larger, more complex molecules. They are the main energy currency of the cell and also play an important role in signalling, both within cells and between them. They are difficult to study because, unlike DNA and proteins, they are not polymers and they behave very heterogeneously. The main technique we use to study them is mass spectrometry, which in a single experiment can detect thousands of signals from different metabolites. The problem is that only a small fraction of those signals can actually be identified – meaning linked to a precise chemical structure. Because metabolites behave unpredictably inside a mass spectrometer, reverse-engineering their structure from the data is extremely difficult. Scientists have been working on this since the 1960s, and yet today, in a typical experiment, most of the data remains unidentified”. To tackle this challenge, you and your team at Princeton develop AI models capable of analysing mass spectrometry data. What are the main results you have achieved so far? “The most exciting result so far has been developing AI models and then experimentally demonstrating that they can actually discover new metabolites. Historically, identifying even a single one could take months, or even a year. The fact that we've now been able to discover more than ninety is remarkable in itself. Looking forward, I hope one day we will comprehensively map all the metabolites in the human body, much like the Human Genome Project mapped all the DNA in the early 2000s. Some of the metabolites we've discovered are structurally very exciting, with chemical elements or functional groups that are unprecedented in mammals. Others appear to be involved in important biological mechanisms: many of the novel metabolites we've found are present in high levels in cancer cells, and two of them decrease very sharply with ageing, both in mice and humans. In collaboration with other biologists, we've been able to work out why these metabolites decline with age and to reverse that process in aged mice”. In which areas of medicine could a more complete map of the human metabolome make a decisive difference? “Right now we are particularly focused on understanding how certain metabolites are associated with a person's risk of developing cancer in the future. There are already several dozen unknown metabolites that various studies have linked to cancer risk. We're very interested in identifying their chemical structures because, in theory, if one could supplement those metabolites or inhibit their production – depending on their effect – that might provide a strategy for cancer prevention, which would be far more effective than cancer treatment”. © Roberto Barnabà / ICTP Photo Archive Your work sits at the intersection of biology, medicine and computational science. How important has your interdisciplinary background been in achieving the results you've just described? “Very important. Throughout my career I've really enjoyed moving into new areas of science and thinking about what the right computational approach to a given problem might be. I think that's led to some genuinely exciting projects and discoveries that I hope will have an impact on human health. It's a somewhat unusual perspective, I'll admit, because science tends still to be conducted in a very disciplinary, siloed way. Having this flexibility and willingness to jump across fields has been a real advantage — it's allowed me to spot unexpected connections. That said, it also has its downsides: I think I have less deep knowledge than I would like in each of the individual fields I work in”. If you were to advise other young scientists, how would you strike the balance between broadening your horizons across different fields and going deep in a single one? “That's a tough one. I think it really depends on where your interests lie. What matters most, in my view, is how much energy and passion you put into your work. If the questions you wake up with in the morning and fall asleep with at night all point to one field, then that's where you should go deep. If different areas pull at you, follow them. There's no point pushing yourself down a path that doesn't genuinely excite you. My advice, ultimately, is to work on the problems you can't stop thinking about”. Atish Dabholkar (ICTP) e Alessandro Curioni (IBM) consegnano il premio a Michael Skinnider. Foto: © Roberto Barnabà / ICTP Photo Archive What does the Brahmagupta Prize for Artificial Intelligence represent for your scientific career so far? “It's an enormous milestone, especially at this stage. Right now my main focus is establishing the reputation of my research group, so this recognition is evidence that the problems we're working on are considered important and that the results we're getting are being noticed. It's also very meaningful on a personal level, precisely because I do such interdisciplinary work. As I mentioned, the downside of that approach is having less specialist knowledge in any single field. So receiving this kind of recognition is a confirmation that this way of doing science can be successful too”. This article was originally written in Italian. Since Il Bo Live does not have native English speakers on staff, we edited the text with the support of an AI-based language tool. L'intervista a Michael Skinnider, che nel suo laboratorio a Princeton sviluppa modelli di intelligenza artificiale capaci di individuare la struttura chimica dei metaboliti. Queste cellule giocano un ruolo decisivo, ma ancora poco conosciuto, per la nostra salute has_gallery NELLA STESSA CATEGORIA Michael A. Skinnider. Foto: Sameer A. Khan / fotobuddy 1 Off