Protein Structure and Dynamics

This paper introduces HydroMap and FastHydroMap models to predict hydrophobicity from water features, enabling rapid and accurate computation of dewetting free energy for proteins and surfaces.

This study uses atomistic simulations to show how ligand binding in one subunit of hGS alters the geometry and solvent environment of the opposite active site, revealing a distributed allosteric network underlying negative cooperativity.

This study uses molecular dynamics simulations to compare unfolding mechanics of 51 keratin proteins, revealing rate-dependent nanomechanical responses and correlations with molecular descriptors.

SLiMNet detects functional similarity in SLiM pairs using deep learning and protein LLM embeddings, identifying new motifs and providing functional annotations.

Promera is a generative model for biomolecular structure prediction, filtering, and design that improves binder filtering and controllable design compared to existing methods.

The paper develops ESM3-PPISites, a model that identifies protein-protein interaction interfaces using multimodal representations, outperforming existing methods and improving docking accuracy.

The paper designs and evaluates D-amino acid-based peptides to inhibit SAA3 fibril formation in a mouse model of cardiac amyloidosis.

This paper benchmarks AI structure predictors on fuzzy protein complexes, finding they consistently violate 30% of NOE restraints and show poor helicity correlation with thermodynamic expectations.

DDI_single improves multi-domain protein structure prediction by accurately assembling domains using a single sequence and a novel gated cross-attention module.

ProtGPT3 is an open-source family of protein language models that enables controlled protein generation through alignment and prompting, showing effectiveness in generating functional proteins.

VelocityFM predicts short-horizon protein trajectories using flow matching in velocity space, achieving high TM-scores and geometric validity on unseen proteins.

EnzOracle predicts enzyme environmental adaptation traits from sequence using a mixture-of-experts framework, showing strong accuracy and mechanism-aware validation.

The paper compares legumain structures across vertebrates, identifying Pocket 2 as a conserved yet variable region with species-specific structural and ligand-binding differences.

The paper improves antibody sequence design by combining a general inverse folding model with an antibody-specific language model, achieving better amino acid recovery and sequence diversity.

This paper introduces FlashABB, a model that predicts antibody structures faster than sequence embedding models, enabling efficient access to 3D information for antibody stability and developability prediction.

This study uses molecular dynamics simulations to investigate how the intrinsically disordered regions of FUS recognize U1 snRNA, revealing distinct interaction mechanisms and structural changes upon RNA binding.

This paper identifies specific interfacial residues in IGPS that modulate allosteric regulation through structural and dynamic changes induced by targeted mutations.

The paper uses language modeling to create a comprehensive representation of protein biology, enabling structure prediction and discovery of high-affinity protein interactions.

AI-IDP predicts conformational ensembles of disordered proteins by combining deep learning with physical restraints, revealing common structural features across proteomes.

The paper maps structural coverage of Arabidopsis thaliana developmental proteins using experimental data and AlphaFold predictions, finding higher structural coverage in plants than in humans.

This paper designs intrinsically disordered proteins using a physics-guided machine learning framework to encode complex signal processing and condensate properties.

The paper determines the atomic structure of insect chorion amyloid filaments and reveals their assembly principles, showing functional amyloids can operate in a regulated biological context.

This study uses a combined experimental platform to analyze the conformational dynamics of EAAT1, revealing that multimodal dynamics regulate its function and identifying a mutation that inhibits transport by stabilizing a paused conformation.

This paper develops a hierarchical sequence-based framework to infer distributed protein organization, identifies novel ciliary proteins, and enables synthetic protein design.

This paper uses pressure and temperature to study how TMV coat protein assembly changes, revealing localized structural dynamics and epitope masking under high pressure.

The paper reports a spontaneous structural transition in the insulin receptor during molecular dynamics simulation, revealing a conformation compatible with hybrid insulin binding sites.

This study uses molecular dynamics simulations to show that antibody maturation increases rigidity in protein-contacting regions and flexibility at glycan interfaces.

This study compares AI tools for predicting protein structural ensembles, finding they generate diverse but distinct conformational distributions, with limited coverage of experimental structures and minimal response to known mutations.

The paper shows that neutral evolution can lead to increased structure in unstructured proteins and loss of structure in structured proteins, suggesting a diffuse sequence-structure landscape.

TAMALE identifies functional residues across the human proteome using structure models and variant effect predictions, validated through case studies and experimental validation.