De novo design of protein structure and function with RFdiffusion.

There has been considerable recent progress in designing new proteins using deep-learning methods1-9. Despite this progress, a general deep-learning framework for protein design that enables solution of a wide range of design challenges, including de novo binder design and design of higher-order symmetric architectures, has yet to be described. Diffusion models10,11 have had considerable success in image and language generative modelling but limited success when applied to protein modelling, probably due to the complexity of protein backbone geometry and sequence-structure relationships. Here we show that by fine-tuning the RoseTTAFold structure prediction network on protein structure denoising tasks, we obtain a generative model of protein backbones that achieves outstanding performance on unconditional and topology-constrained protein monomer design, protein binder design, symmetric oligomer design, enzyme active site scaffolding and symmetric motif scaffolding for therapeutic and metal-binding protein design. We demonstrate the power and generality of the method, called RoseTTAFold diffusion (RFdiffusion), by experimentally characterizing the structures and functions of hundreds of designed symmetric assemblies, metal-binding proteins and protein binders. The accuracy of RFdiffusion is confirmed by the cryogenic electron microscopy structure of a designed binder in complex with influenza haemagglutinin that is nearly identical to the design model. In a manner analogous to networks that produce images from user-specified inputs, RFdiffusion enables the design of diverse functional proteins from simple molecular specifications.

A dynamic simulation model to support reduction in illegal trade within legal wildlife markets.

Sustainable wildlife trade is critical for biodiversity conservation, livelihoods, and food security. Regulatory frameworks are needed to secure these diverse benefits of sustainable wildlife trade. However, regulations limiting trade can backfire, sparking illegal trade if demand is not met by legal trade alone. Assessing how regulations affect wildlife market participants' incentives is key to controlling illegal trade. Although much research has assessed how incentives at both the harvester and consumer ends of markets are affected by regulations, little has been done to understand the incentives of traders (i.e., intermediaries). We built a dynamic simulation model to support reduction in illegal wildlife trade within legal markets by focusing on incentives traders face to trade legal or illegal products. We used an Approximate Bayesian Computation approach to infer illegal trading dynamics and parameters that might be unknown (e.g., price of illegal products). We showcased the utility of the approach with a small-scale fishery case study in Chile, where we disentangled within-year dynamics of legal and illegal trading and found that the majority (∼77%) of traded fish is illegal. We utilized the model to assess the effect of policy interventions to improve the fishery's sustainability and explore the trade-offs between ecological, economic, and social goals. Scenario simulations showed that even significant increases (over 200%) in parameters proxying for policy interventions enabled only moderate improvements in ecological and social sustainability of the fishery at substantial economic cost. These results expose how unbalanced trader incentives are toward trading illegal over legal products in this fishery. Our model provides a novel tool for promoting sustainable wildlife trade in data-limited settings, which explicitly considers traders as critical players in wildlife markets. Sustainable wildlife trade requires incentivizing legal over illegal wildlife trade and consideration of the social, ecological, and economic impacts of interventions.

Un Modelo Dinámico de Simulación para Asistir en la Reducción del Comercio Ilegal dentro de Mercados Legales de Vida Silvestre Resumen El comercio sustentable de vida silvestre es crítico para la conservación de la biodiversidad, los medios de subsistencia y la seguridad alimentaria. Son necesarios marcos regulatorios para asegurar estos diversos beneficios del comercio sustentable de vida silvestre. Sin embargo, las regulaciones que limitan el comercio pueden ser contraproducentes, generando un mercado ilegal si la demanda no se suple solamente con el comercio legal. El análisis de cómo las regulaciones afectan a los incentivos de los participantes del comercio de vida silvestre es de suma importancia para controlar el comercio ilegal. Mientras que muchas investigaciones se han centrado en analizar cómo las regulaciones afectan tanto a quienes consumen como quieren proveen visa silvestre, , poco se ha hecho para entender los incentivos de los intermediarios. Construimos un modelo dinámico de simulación para asistir en la reducción del comercio ilegal de vida silvestre dentro de los mercados legales, enfocándonos en los incentivos que enfrentan los intermediarios para comercializar productos legales o ilegales. Usamos un enfoque de Computación Bayesiana Aproximada para inferir las dinámicas del comercio ilegal y los parámetros que podrían ser desconocidos (p. ej.: el precio de los productos ilegales). Demostramos la utilidad del modelo mediante el caso de estudio de una pesquería de pequeña escala en Chile, en donde desentrañamos las dinámicas del comercio legal e ilegal y estimamos que la mayor parte del pescado comercializado es ilegal. Utilizamos el modelo para analizar el efecto de intervenciones para mejorar la sustentabilidad de la pesquería y para explorar los trade-offs entre metas ecológicas, económicas y sociales. Las simulaciones de escenarios mostraron que incluso incrementos significativos (más del 200%) de parámetros que recreaban intervenciones permitieron solamente mejoras moderadas en la sustentabilidad ecológica y social de la pesquería a un costo económico sustancial. Estos resultados exponen cuán desequilibrados están los incentivos de los intermediarios hacia el comercio de productos ilegales por encima de los legales en esta pesquería. Nuestro modelo proporciona una herramienta innovadora para la promoción del comercio sustentable de vida silvestre en entornos con datos limitados, y considera explícitamente a los intermediarios como actores críticos dentro del comercio de vida silvestre. El comercio sustentable de vida silvestre requiere incentivar el comercio legal sobre el ilegal y la consideración del impacto social, ecológico y económico de las intervenciones.

Improved motif-scaffolding with SE(3) flow matching.

Protein design often begins with knowledge of a desired function from a motif which motif-scaffolding aims to construct a functional protein around. Recently, generative models have achieved breakthrough success in designing scaffolds for a diverse range of motifs. However, the generated scaffolds tend to lack structural diversity, which can hinder success in wet-lab validation. In this work, we extend FrameFlow, an SE(3) flow matching model for protein backbone generation, to perform motif-scaffolding with two complementary approaches. The first is motif amortization, in which FrameFlow is trained with the motif as input using a data augmentation strategy. The second is motif guidance, which performs scaffolding using an estimate of the conditional score from FrameFlow, and requires no additional training. Both approaches achieve an equivalent or higher success rate than previous state-of-the-art methods, with 2.5 times more structurally diverse scaffolds. Code: https://github.com/microsoft/frame-flow.