Jianyi Yang

RiboDFM designs RNA sequences that fold into target 3D structures using a backbone-conditioned discrete flow-matching framework, achieving high sequence recovery and tertiary-structure consistency.

CryoCodex improves cryo-EM map quality by jointly enhancing density, estimating local quality, and masking molecules, outperforming existing methods on experimental maps.

trFlow generates all-atom protein conformations faster and more accurately than existing methods by using flow matching with three key innovations.

CryoForge is an evidence-gated agent that improves cryo-EM model building by selectively accepting or rejecting edits based on structural evidence, achieving validated improvements in most test cases.

This paper introduces trRosettaRNA2, a deep learning model for RNA 3D structure prediction that outperforms existing methods with fewer parameters, and demonstrates its effectiveness on RNase P RNA.

The paper benchmarks MSA pairing strategies for protein-protein complex structure prediction and finds that increased MSA depth, not specific pairing, improves accuracy.

RNArank assesses RNA 3D structure model quality using deep learning and intermediate 2D maps, outperforming existing methods in benchmark tests.

This paper introduces trX2-D, a method that combines deep learning with physics-based sampling to predict protein dynamic structures and alternative conformations.

This study identifies the Archaeal Ribosome regulatory Complex (ARC) as a dual-factor complex that modulates ribosome hibernation in response to energy status.

PathDiffusion models protein folding pathways using evolution-guided diffusion, accurately reconstructing folding events and demonstrating versatility across multiple applications.

The paper introduces mTM-align2, a method for rapid and accurate protein structure database search that uses inverse folding models and contrastive learning, achieving high precision in seconds.

The paper introduces SaprotHub, a platform that simplifies training and sharing of protein language models, enabling non-experts to build and collaborate on customized models.

CryFold improves de novo model building for cryo-EM maps using local attention and 3D rotary position embedding, producing more complete models and reducing resolution requirements.

This paper introduces IDPFold, a diffusion model that generates conformational ensembles for intrinsically disordered proteins directly from sequences, without requiring MSA or experimental data.

PLMSearch uses a protein language model to perform fast and accurate homologous protein sequence search, improving sensitivity over traditional methods while matching structure search performance.

Q-BioLiP improves upon BioLiP by representing protein-ligand interactions using quaternary structures and providing both experimental and predicted binding affinities.

PGAR-Zernike provides a fast and accurate open-source toolkit for computing Zernike descriptors of protein structures, enabling efficient structural database construction and retrieval.

This paper introduces trRosettaX-Single, a single-sequence protein structure prediction method that outperforms AlphaFold2 and RoseTTAFold on natural and designed proteins.

This paper introduces trRosettaRNA, a deep learning method for RNA 3D structure prediction, showing improved accuracy over existing methods on benchmark datasets.

This study identifies SARS-CoV-2 N protein interactions with host proteins and mRNAs, showing it attenuates stress granule formation and alters gene expression.

The paper identifies recombination regions and mutation sites in SARS-CoV-2, SARS-CoV, and SARS-like CoVs, linking specific genetic features to pandemic emergence.

This paper introduces a deep residual network and Rosetta protocol for predicting inter-residue orientations and distances, outperforming existing methods in benchmark tests.