Deep generative models are increasingly powerful tools for the in silico design of novel proteins. Recently, a family of generative models called diffusion models has demonstrated the ability to generate biologically plausible proteins that are dissimilar to any actual proteins seen in nature, enabling unprecedented capability and control in de novo protein design. However, current state-of-the-art diffusion models generate protein structures, which limits the scope of their training data and restricts generations to a small and biased subset of protein design space. Here, we introduce a general-purpose diffusion framework, EvoDiff, that combines evolutionary-scale data with the distinct conditioning capabilities of diffusion models for controllable protein generation in sequence space. EvoDiff generates high-fidelity, diverse, and structurally-plausible proteins that cover natural sequence and functional space. We show experimentally that EvoDiff generations express, fold, and exhibit expected secondary structure elements. Critically, EvoDiff can generate proteins inaccessible to structure-based models, such as those with disordered regions, while maintaining the ability to design scaffolds for functional structural motifs. We validate the universality of our sequence-based formulation by experimentally characterizing intrinsically-disordered mitochondrial targeting signals, metal-binding proteins, and protein binders designed using EvoDiff. We envision that EvoDiff will expand capabilities in protein engineering beyond the structure-function paradigm toward programmable, sequence-first design.
17 September 2018 in Physical Chemistry Chemical Physics
Type III antifreeze proteins (AFP III) have been widely recognized as one class of ice-binding proteins produced by several biological organisms to withstand freezing conditions. Besides their ability to restrict ice growth through their ice-binding site (IBS), AFP III have also been shown to possess a great propensity for hydrophobic surfaces such as the air–water interface. Yet, it is not known whether AFP III adsorb with a specific orientation and how hydrophobic interactions affect the IBS. Molecular insights on the accessibility of the IBS and its interactions with water are important for understanding AFP III action in vivo but also for their application as ice-inhibiting agents for deicing, frozen food storage, as well as for long-term blood and organ cryo-preservation. Here, the orientation of fish AFP III adsorbed at the air–water interface has been studied using a combination of molecular dynamics (MD) simulations and vibrational sum-frequency generation (SFG) spectroscopy together with spectral calculations. The SFG/MD analysis indicated that when AFP III adsorbs at the air–water interface, it mostly retains its native state and orients with a tilt angle of 120° with respect to the surface normal. We found that the IBS is only partially solvated, leaving the pyramidal ice plane binding domain exposed to the vapor phase. These findings suggest that interactions with hydrophobic interfaces (e.g., cell membranes, polymers) could lead to the partial decoupling of the IBS from water and, to some extent, to a loss of AFP III antifreezing activity.
Modern biology is powered by the organization of biological information, a framework pioneered in 1965 by Margaret Dayhoff's Atlas of Protein Sequence and Structure. Databases descended from this common ancestor power computational methods that have revolutionized our ability to understand and design proteins. We introduce the Dayhoff Atlas, a centralized collection of both protein sequence data and generative protein language models, as a modern-day resource for protein biology in the age of AI. We combined and reclustered metagenomic sequences with UniRef100 to create GigaRef, which includes 3.34 billion protein sequences across 1.70 billion clusters and provides the largest open dataset of natural proteins to date. To fuse the richness of protein structure with the scalability of sequence space, we generated structure-based synthetic data, producing BackboneRef, a first-in-class dataset of 46 million synthetic protein sequences predicted from 240,811 de novo designed backbones. Using these data and evolutionary sequence alignments, we trained the Dayhoff family of protein language models, including the first model that combines single proteins and sets of evolutionarily-related sequences at scale. Operating only in amino acid sequence space, the Dayhoff models can natively predict mutation effects on fitness, scaffold structural motifs, and perform guided generation of new proteins within a specific family. Learning from metagenomic and structure-based synthetic data from the Dayhoff Atlas increased the cellular expression rates of generated proteins, highlighting the real-world value of expanding the scale, diversity, and novelty of protein sequence data. We release the Dayhoff Atlas, including all datasets, code, and models, under a permissive license to continue to bring the power of computation to protein biology and design.
20 February 2025 in Current Opinion in Structural Biology