LAUSR

Significance Protein sequences change over evolutionary times. The fixation of these changes is coupled to the protein’s capability to properly explore conformational space and perform biological functions. Repeat proteins are privileged systems to understand the relations between evolution and folding, due to their conserved structure and functional diversity. Here, we use the evolutionary record of natural sequences to model and analyze the folding mechanisms of thousands of proteins. The proposed model successfully reproduces folding experiments using only sequence information as input. We performed large-scale predictions of folding mechanisms in the most abundant repeat-protein family, identifying higher-order features such as domain emergence, stability, and cooperativity of repeat arrays.