Skip to main content
Environment

A structure-aware generative AI framework for revealing functional relationships in protein families.

| Source: Proceedings of the National Academy of Sciences of the United States of America

Proteins can be studied through their sequence statistics or structural properties. These represent complementary views that are useful but lack a quantitative framework to tell, family by family, which is most informative and how to combine them. We introduce a framework that builds both views in parallel: amino acid (AA) alignments are translated into parallel alignments over a 3D interaction (3Di) structure-informed alphabet. Variational autoencoders compress each into a two-dimensional map,

Proteins can be studied through their sequence statistics or structural properties. These represent complementary views that are useful but lack a quantitative framework to tell, family by family, which is most informative and how to combine them. We introduce a framework that builds both views in parallel: amino acid (AA) alignments are translated into parallel alignments over a 3D interaction (3Di) structure-informed alphabet. Variational autoencoders compress each into a two-dimensional map, and direct coupling analysis places a shared coevolutionary energy on both maps, turning them into latent generative landscapes. On these landscapes, we define information-theoretic distance metrics that quantify how sequence changes drive structural and functional variation in protein families. We demonstrate the framework on five families: in malate dehydrogenases, the 3Di landscape identifies the structurally conserved scaffold that this family uses to encode thermal adaptation via sequence variability revealed in the AA landscape. In globins and transient receptor potential melastatin (TRPM), the 3Di landscape recovers known functional subfamilies. In the Flaviviridae E1 and E2 glycoproteins, structure reveals evolutionary relationships invisible at the sequence level. Because many sequences encode the same fold, our framework lets us disentangle family-sequence variability from structural and functional variation. These generative landscapes allow sampling near functional regions, and we show they can help us gain mechanistic insight into the evolutionary forces shaping sequence-structure-function variation and guide the design of new proteins.

Read the original source →

Related Stories

Environment

Mussels possess distinct vanadium and iron storage systems for fabricating metal-reinforced adhesive fibers.

Mussels fashion a protein-based fibrous adhesive holdfast known as a byssus to tether themselves in marine habitats against crashing waves. The capacity of mussels to anchor securely is dependent on load-bearing protein-metal cross-links mediated by 3,4-dihydroxyphenylalanine (DOPA) residues coordinated to metal ions, which significantly enhance byssus mechanical properties (e.g., adhesive strength, abrasion resistance, toughness, self-healing capacity). However, the natural metal incorporation

Continue reading
Environment

Host and anatomical filters prevent microbiome convergence in avian scavengers.

How host-associated microbiomes assemble under intense environmental microbial exposure remains a central question in microbial ecology. Vultures represent a unique natural system because these species differ markedly in feeding ecology and carcass use. We characterized bacterial communities across four anatomical niches (facial, buccal, tracheal, and cloacal) in 155 individuals spanning all four European obligate vulture species (subfamily Gypaetinae: Gypaetus barbatus, Neophron percnopterus, a

Continue reading
Environment

Physical distancing to prevent influenza-like illness on a college campus: The eX-FLU trial.

Public health recommendations often advise people to isolate when infected with influenza-like illness (ILI) with the goal of promoting physical distancing and limiting face-to-face interactions when risk of transmission is highest. Because isolation may have negative consequences, it is important to carefully assess whether encouragement to physically distance actually prevents cases of ILI. College campuses can be especially vulnerable to outbreaks of ILI and are ideal environments for assessi

Continue reading
Environment

Dynamical geometric modes in non-Euclidean plates.

Continuum elastic shells with residual stresses can exhibit complex motions that require vanishing elastic energy to excite, also called geometric zero modes, enabling them to be driven by weak and generic energy inputs. Despite recent interest in these modes, we understand very little about their dynamical properties. Plates with a curved, rather than flat, in-plane geometry, also called non-Euclidean plates, are one example in which residual stresses and geometry combine to produce a continuum

Continue reading
Environment

Hadean mantle depletion and 3.95-Gyr subduction implied by North China rocks.

The scarcity of terrestrial rocks older than 3.9 billion years (Gyr) has hindered our understanding of Earth's early evolution, fuelling persistent debates on how and when the first continental crust formed and plate tectonics started. At present, the only unequivocally documented >3.9-Gyr-old rocks are from the Acasta Gneiss Complex, Canada 1-3 . Here we report a 3.95-Gyr-old trondhjemitic gneiss from eastern Hebei, North China Craton, in which primary magmatic zircon records elevated

Continue reading