Evolutionary Intelligence: How Coevolution Opened the Way to the AlphaFold Revolution
Year: 2026
Type: article
Abstract: The sequences and structures of proteins that exist in nature are a joint product of physical mechanisms and evolutionary processes that are deeply interconnected. It was proposed that physical interactions between amino acids at different protein sites induce evolutionary correlations (co-evolution) between these sites. The inference of these correlations, first as mutual information, then as parameters of a statistical model that describes protein families (direct couplings) and, finally, through the deep neural networks of AlphaFold, produced very powerful developments that revolutionized the field of protein structure and function prediction, providing unprecedented advances in the understanding of proteins and even cells. To discuss the evolutionary mechanisms that produce the correlations exploited by AlphaFold, some of us organized the symposium “Evolutionary intelligence: How coevolution opened the way to the AlphaFold revolution” (Madrid, August 2025), sponsored by the Journal of Molecular Evolution and by the Computational Biology and Bioinformatics (BCB) Hub of the Spanish Research Council (CSIC). The symposium gathered a collection of researchers who contributed to the development and the advancement of the approach based on evolution-induced correlations. This paper represents a collective discussion of part of this community.
Source: Journal of Molecular Evolution
Institutions: Centro de Biología Molecular Severo Ochoa, Universidad Autónoma de Madrid, KTH Royal Institute of Technology, Centre National de la Recherche Scientifique, École Normale Supérieure - PSL, Sorbonne Université, École Normale Supérieure, Temple University, Scuola Internazionale Superiore di Studi Avanzati, Karolinska Institutet, Centro Andaluz de Biología del Desarrollo, Junta de Andalucía, Universidad Pablo de Olavide, University of Manchester, Institució Catalana de Recerca i Estudis Avançats, Barcelona Supercomputing Center
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: hybrid
Estimated APC paid: $4,090
OpenAlex ID: W7225231328