Breaking: The official release of educational guides for AlphaFold Server from @GoogleDeepMind. These tutorials aim to assist new users in gaining a deeper understanding of the latest AlphaFold version’s capabilities and maximizing the potential of

AF3Complex Yields Improved Structural Predictions of Protein Complexes. １．AF3Complex is a new model derived from AlphaFold 3 that improves the prediction of protein complex structures, especially interfacial accuracy, surpassing AlphaFold 3 in various benchmark settings.

AF3Complex Yields Improved Structural Predictions of Protein Complexes. １．AF3Complex is a new model derived from AlphaFold 3 that improves the prediction of protein complex structures, especially interfacial accuracy, surpassing AlphaFold 3 in various benchmark settings.

Data-driven protease engineering by DNA-recording and epistasis-aware machine learning @NatureComms. １．This study introduces a high-throughput DNA recorder system that enables deep specificity profiling of proteases by linking protease-substrate activity to DNA recombination

Data-driven protease engineering by DNA-recording and epistasis-aware machine learning @NatureComms . １．This study introduces a high-throughput DNA recorder system that enables deep specificity profiling of proteases by linking protease-substrate activity to DNA recombination

LassoPred: a tool to predict the 3D structure of lasso peptides @NatureComms. １．LassoPred is a specialized structure prediction tool built to tackle the unique challenges of lasso peptides (LaPs), which feature a slipknot-like [1]rotaxane topology with an isopeptide-bonded ring

LassoPred: a tool to predict the 3D structure of lasso peptides @NatureComms . １．LassoPred is a specialized structure prediction tool built to tackle the unique challenges of lasso peptides (LaPs), which feature a slipknot-like [1]rotaxane topology with an isopeptide-bonded

AlphaFold distillation for inverse protein design @SciReports. １．A new approach distills AlphaFold’s structure confidence metrics (pTM, pLDDT) into a lightweight model, AFDistill, enabling fast and differentiable structural scoring directly from protein sequences — no structure

AlphaFold distillation for inverse protein design @SciReports . １．A new approach distills AlphaFold’s structure confidence metrics (pTM, pLDDT) into a lightweight model, AFDistill, enabling fast and differentiable structural scoring directly from protein sequences — no

Comprehensive datasets for RNA design, machine learning, and beyond @SciReports. １．A new benchmark dataset of over 320,000 RNA secondary structures has been released, addressing a critical need in the RNA design and modeling community. It enables more robust training and

Comprehensive datasets for RNA design, machine learning, and beyond @SciReports . １．A new benchmark dataset of over 320,000 RNA secondary structures has been released, addressing a critical need in the RNA design and modeling community. It enables more robust training and

Large-scale predictions of alternative protein conformations by AlphaFold2-based sequence association @NatureComms. １．A new method called CF-random, based on AlphaFold2 and ColabFold, efficiently predicts alternative protein conformations—including fold-switching, rigid body

Large-scale predictions of alternative protein conformations by AlphaFold2-based sequence association @NatureComms . １．A new method called CF-random, based on AlphaFold2 and ColabFold, efficiently predicts alternative protein conformations—including fold-switching, rigid body

In-silico 3D molecular editing through physics-informed and preference-aligned generative foundation models @NatureComms . １．The paper introduces MolEdit, a generative foundation model for 3D molecular editing, which integrates physical laws and user-specified preferences into

In-silico 3D molecular editing through physics-informed and preference-aligned generative foundation models @NatureComms. １．The paper introduces MolEdit, a generative foundation model for 3D molecular editing, which integrates physical laws and user-specified preferences into

SE(3)-equivariant ternary complex prediction towards target protein degradation @NatureComms. １．DeepTernary is a new deep learning model that predicts ternary complex structures formed by small molecules (like PROTACs or MGDs), a protein of interest (POI), and an E3 ligase.

SE(3)-equivariant ternary complex prediction towards target protein degradation @NatureComms . １．DeepTernary is a new deep learning model that predicts ternary complex structures formed by small molecules (like PROTACs or MGDs), a protein of interest (POI), and an E3 ligase.

A generalized platform for artificial intelligence-powered autonomous enzyme engineering @NatureComms. １．Researchers present a general-purpose, autonomous enzyme engineering platform that integrates protein language models, machine learning (ML), and robotic biofoundry

A generalized platform for artificial intelligence-powered autonomous enzyme engineering @NatureComms . １．Researchers present a general-purpose, autonomous enzyme engineering platform that integrates protein language models, machine learning (ML), and robotic biofoundry