Studying the genome? Like the first trip into space
"Today we have very advanced computational techniques available, but to solve complex problems it is necessary to create and teach a new approach which must be multidisciplinary by necessity". Francesca Buffa, a new entry among the tenured faculty of the Computing Sciences department, explains the ultimate reason for the intellectual melting pot that is giving life to the latest Bocconi research outfit. She graduated in Theoretical Physics from the University of Turin, specializing in computational sciences, and until a few months ago she was Professor of Computational Biology and Cancer Genomics at Oxford. Her work is the demonstration of the extent to which artificial intelligence and machine learning can be integrated with the physical, natural and social sciences to open new frontiers of research.
When you chose to study physics, would you ever have imagined working together with doctors and biologists?
Maybe not when I was at university, but the application of my studies to medical sciences is not something that happened to me by chance, it is a path that I chose at a precise moment, which I remember very well. When the sequencing of the human genome was published in Cambridge in 2001, I was in London for a PhD in Applied Mathematics and, like other scientists, I felt an instinctive attraction for this new world. For us it was like the first trip to the moon, the discovery of an unexplored universe and a frontier of knowledge that opened up before our eyes. It was a very emotional moment and I immediately decided that this would be my next challenge.
In your most recent research, which you began in Oxford and will continue at Bocconi, the techniques of computational science are applied, for example, to the treatment of glioblastoma. What are the prospects and objectives?
One of the main obstacles to the development of new therapies for this aggressive form of brain tumor is represented by the scarcity of biomarkers to guide the development of new drugs. The idea is therefore to follow the patients before and after the therapies, which unfortunately often are not completely effective, by measuring the RNA circulating in the blood and trying to understand the relationship between ionizing radiation, tumor progression and the genome of patients. These are complex data, which involve numerous variables and must also be integrated with imaging data from MRI or PET scans. Machine learning methods are indispensable at all stages of the analysis, to map these data, clean them up, understand and integrate them, and finally arrive at a kind of modeling that helps us predict which patients will respond to the treatment.
The value of such a branched and multidisciplinary study, therefore, is not only realized at its conclusion...
The understanding of these data takes place in progressive steps and by bringing together knowledge that comes from different scientific environments and biomedical applications. For example, we can start from information coming from cell cultures using new techniques such as CRISPR that allows us to "turn off" genes one at a time and measure the cellular reaction. We can then measure these reactions and understand in more detail how cells change. There are numerous biobanks that combine data and genetic profiles of healthy individuals to understand how likely they are to develop certain diseases and possibly activate large-scale screening programs without necessarily sequencing the entire genome of individuals. But it is also possible to build computational models to understand the stages of the development of a disease, to simulate its progression or the response to a drug treatment. The results we see in modern medicine, in all fields, including vaccines, would not be possible, especially in such a compressed timespan, without the help of computing sciences and without a scientific community composed of computer scientists, biologists, physicists, mathematicians and doctors willing to work together. Today knowledge can travel in all possible directions as never before.
How is all this taught to the younger generations?
In my training I have been in contact with very diverse areas and modes of study and I try to replicate this wide view by creating occasions in the classroom that are set up as discussions where there is dialogue between as many voices as possible. My first taste of teaching at Bocconi with the AI ​​Lab put me in contact with a group of students who were very prepared and very eager, and for whom this open approach to contamination seemed absolutely natural. None of them were in the classroom just to learn techniques, they all posed questions that went beyond specific examples and animated discussions that sometimes went to a very high level.
Women and STEM subjects: is there really a complicated relationship between the two? What situation did you find in the English or Japanese universities you attended?
The problem of a weaker female presence in STEM degrees is everywhere. In countries such as England, where the gender imbalance on scientific degrees is less than in Italy, there is an equally strong gender gap between most senior positions, such as department heads. To achieve a cultural change that allows girls to dive more courageously into these disciplines, it is necessary to renew primary education in elementary and middle schools. If I think about my experience, I consider myself very lucky, not only because I had excellent teachers but because I had parents who supported me in my studies and helped me to be free to make my choices. It was my mother, a philosopher, who actually encouraged me to pursue my passion for science and physics!
Francesca Buffa is one of the new full professors of the newly established Department of Computing Sciences, where she coordinates the Artificial Intelligence laboratory. Internationally renowned researcher, she comes from the University of Oxford, where she was Professor of Computational Biology and Cancer Genomics. She is Principal Investigator of an ERC Consolidator Grant for the development of computational methods for the study of complex diseases, a grant from the Invernizzi Foundation awarded in the context of an ERC Call by the Cariplo Foundation, and a very recent grant from the Celeghin Foundation. The latter also involves Pisa University Hospital and the Pisa Foundation for Science in the study of the biomarkers of patients suffering from glioblastoma, one of the most aggressive brain tumors. "At Bocconi I have found an international environment and a speed in organizational systems that are at a par with the best Anglo-Saxon universities. In these weeks I have never felt out of place: I personally experienced the strong determination to work on important issues and a great ability to open up to new things and integrate them. I'm sure Milan will be equally welcoming to my family: my husband and our three children are moving here from Oxford to join me soon."