Nov. 29, 2023 — Artificial intelligence and new computational methods are revolutionizing scientific research. One of the fields where their influence and impact is most evident is in cancer, where the amount of available data is constantly accumulating.
To explore and share the latest advances in these areas, the Barcelona Supercomputing Center – Centro Nacional de Supercomputación (BSC-CNS) and Columbia University, with the support of the Asociación Española Contra el Cáncer (AECC), organized the symposium “AI Meets Cancer”, being held on November 29 and 30 at the Campus Nord of the Universitat Politècnica de Catalunya (UPC) in Barcelona.
“The rapid evolution of artificial intelligence, not only of ChatGPT at a popular level but in all scientific applications, makes the organization of this meeting more than justified, which will also establish a stable collaboration between these institutions,” said Alfonso Valencia, director of the Department of Life Sciences at the BSC and co-organizer of the symposium.
“On the one hand, recent technological developments in biomedicine are generating a large amount of data. On the other, new computational techniques – such as Large Language Models – allow us to extract knowledge from them and establish predictive models that go beyond the descriptive models to which we are traditionally accustomed,” explained Raúl Rabadán, director of the Mathematical Genomics Program at Columbia University and co-organizer of the event. “The purpose of this meeting is to bring experts from both fields to establish points of interest, encourage interaction and accelerate the inevitable process of integration,” he stresses.
For Fátima Al-Shahrour, head of the Bioinformatics Unit at the Centro Nacional de Investigaciones Oncológicas (CNIO) and one of the speakers at the conference, it is “an innovative meeting that allows different institutions and groups, dedicated both to cancer research from a bioinformatics perspective and to artificial intelligence and computational modelling, to join forces and share challenges.”
A Compendium of Sessions and Topics
The symposium is structured around four main themes. The first will focus on ways of analyzing cancer data, with a special interest in predicting responses to treatments and on single cell analysis methods, which facilitate understanding the origin and development of a tumor, as well as the possible and feared relapses, originating from “cells that are probably present from the beginning, which resist treatment and proliferate after it,” explained Valencia.
The second axis will focus on new technologies and reflect on new applications. For example, language models similar in nature to ChatGPT, which can also be useful in cancer data analysis. The third part will bring artificial intelligence down to the molecular level: it will discuss how it can help predict the influence of different mutations and discern which ones will have an impact on the biology of each tumor. Finally, “specific cases and specific studies that combine all of the above will be presented,” explained Valencia.
“The role of artificial intelligence is going to be very important, not only in research, but also in the treatment and prevention of cancer and other diseases,” said Rabadán. “In research, for example, it will allow us to do virtual experiments: what happens if we change an amino acid or introduce a mutation we find in a patient? How do mutations inherited from our parents contribute to the risk of certain tumors? How can we optimize personalized therapies taking into account the molecular and clinical data of each patient? These are just a few of many questions. And we are only at the beginning,” he said.
The Role of Supercomputing
The event is aligned with the start-up phase at the BSC of the MareNostrum 5, which will be the third most powerful supercomputer in Europe and the eighth in the world to advance research in artificial intelligence. “In any development of artificial intelligence, supercomputing capacity is essential. Having our own infrastructures that support centres and institutions is essential to avoid creating bottlenecks in research,” said Al-Shahrour.
“Cancer is the area of biology in which computation is most advanced,” said Valencia. “The BSC was already involved in the largest such project to date, the International Cancer Genome Consortium (ICGC), in which thousands of complete cancer genomes were analysed for the first time, and we have always been interested in putting large computational capabilities at the service of cancer research. The new challenges posed by artificial intelligence make the need for large facilities like the BSC even more obvious.”
Other participants at the meeting include: Mohammed Alquraishi (Columbia University Medical Center); María Rodríguez Martínez (Zurich Research Laboratory, IBM); Julio Sáez Rodríguez (EMBL – University Hospital Heidelberg); Charlotte Bunne (ETH Zurich); David Jones (University College London).
Source: Barcelona Supercomputing Center – Centro Nacional de Supercomputación