What do the 2024 Nobel Prizes in Physics and Chemistry have in common with the CoMPaSS-NMD Project?

Two words: Artificial Intelligence (AI).

“For fundamental discoveries and inventions that have made machine learning possible with artificial neural networks”. This is the official recognition given to the 2024 Nobel Prize in Physics, jointly awarded to John Hopfield and Geoffrey Hinton.

The 2024 Nobel Prize in Chemistry concerns proteins. It is shared by David Baker, who succeeded in the almost impossible task of building new types of proteins, and Demis Hassabis and John Jumper who developed an artificial intelligence model capable of predicting the complex three-dimensional structures of proteins, thus solving a 50-year-old problem.

CoMPaSS-NMD, a project funded by the European Union under the Horizon Europe Program and coordinated by the University of Modena and Reggio Emilia, will use artificial intelligence to develop more precise and rapid diagnoses of hereditary neuromuscular diseases.

An “Alien” Intelligence

“Sometimes I think it’s as if aliens had landed but people hadn’t realized it because they speak good English”. These are Hinton’s words about Artificial Intelligence.

Let’s try to understand this “alien intelligence” better.

Artificial intelligence models are based on data, that is, on information that feeds and trains these models, which they then use to process and reprocess the data to produce a result.

If data feeds artificial intelligence, what makes it work is the algorithm, or rather the instructions that, through programming, we define for a computer system.

Artificial intelligence algorithms are based on machine learning, of which deep learning, which earned Hopfield and Hinton the Nobel Prize, is a particular type.

Through various techniques, machine learning allows machines to learn from historical data without being explicitly programmed for each task. Deep learning, on the other hand, relies on artificial neural networks, inspired by the functioning of the human brain, to process complex data and build sophisticated models.

The explosion of machine learning in recent years has been made possible by two main factors: data availability and computing power.

The enormous amount of data generated every day feeds learning algorithms, providing examples from which to learn and improve constantly, and this allows them to become increasingly sophisticated.

The increase in computer power has made it possible to train increasingly large and complex neural networks, with billions of parameters.

The Interdisciplinarity of Artificial Intelligence

The 2024 Nobel Prize in Physics, which “deservedly” rewards the pioneers of artificial intelligence, “is proof of how Physics is a discipline with increasingly fluid boundaries and in continuous expansion”,  says Giorgio Parisi, Nobel Prize in Physics in 2021.

Machine learning, in fact, has found applications in numerous fields, from physics to biology, to medicine. In particular, it has revolutionized the way scientists analyze data, enabling discoveries that were previously unthinkable. Some examples include:

  • Particle physics: discovery of the Higgs boson and analysis of gravitational waves.
  • Materials science: design of new materials with desired properties.
  • Biochemistry: study of proteins and discovery of new drugs.
  • Medicine: patient care, telemedicine, pharmacovigilance and accurate diagnosis and personalized treatments.

AI in Medicine: the CoMPaSS-NMD Project

The CoMPaSS-NMD project is precisely about the application of artificial intelligence in medicine. In this project, artificial intelligence will support and guide doctors towards a more accurate diagnosis of neuromuscular diseases (HNMD).

Currently, the diagnosis of these diseases is complex due to the similarity of symptoms between different neuromuscular diseases.

CoMPaSS-NMD uses AI to overcome these difficulties and offer a more accurate diagnosis, thus improving the quality of life of patients and their caregivers.

The AI-based systems that CoMPaSS-NMD is working on integrate and analyze various sources of data, including genetic information, clinical datasets, magnetic resonance imaging and muscle histology data, and information reported by patients.

Computers can process, integrate, and analyze a large amount of data collected from hundreds of patients from different European centers and group them into clusters that have similar clinical characteristics. The more data they have, the more efficient and accurate they become in analyzing and interpreting it. The AI model is able to identify relationships between data that are not recognizable with standard analytical methods.

The CoMPaSS-NMD study is divided into two phases.

A retrospective observational study conducted on genetic, histopathological, and MRI data will use existing data collected from centers, partners of the CoMPaSS-NMD consortium, in the UK, France, Finland, and Italy. This data is used to train machine learning.

A prospective study involves the collection of new clinical, genomic, histopathological, and MRI data obtained from hundreds of patients with neuromuscular diseases who have not received a diagnosis. This data, coming from clinical centers in Italy and Germany, is used to test and validate the machine learning algorithms trained in the previous phase of the study.

In addition to a new diagnostic approach based on AI, one of the main results of the project is the creation of a Neuromuscular Genome Atlas, the CoMPaSS-NMD Atlas: an archive of genetic data, muscle magnetic resonance imaging, and tissue analysis of HNMD patients from the six consortium clinical centers, which will encourage the development of strategies that integrate health data to support patients, healthcare professionals, and citizens.

AI is now a transformative technology that is revolutionizing numerous sectors, including medicine. The combination of large amounts of data and computing power is making it possible to develop increasingly sophisticated models, with a significant impact on scientific research and society in general. And CoMPaSS-NMD goes precisely in this direction: allowing a faster and more accurate diagnosis of HNMD and promoting effective actions by European national health systems, improving the quality of life of patients and those who care for the sick, and reducing costs for both patients and the healthcare system.