Deep-learning models have transformed our ability to predict the phenotypic effects of sequence perturbations. Protein language models (pLMs) score the evolutionary propensity of any amino-acid substi...

Protein sequences are constrained not only by the need to fold into stable structures, but also by specific functional requirements imposed by natural selection. Yet predictions of how amino-acid chan...

Vota a las finalistas del premio Vanguardia de la Ciencia

Misspecification is a blessing, not a curse, when estimating protein fitness from evolutionary sequence data using generative models.

Beyond Scaling: What Truly Works in Protein Fitness Prediction

Generative models trained on natural sequences are increasingly used to predict the effects of genetic variation, enabling progress in therapeutic design, disease risk prediction, and synthetic biolog...