Daisuke Kihara

CryoPilot uses two language model agents to automate cryo-EM structure determination, improving resolution and model accuracy through adaptive decision-making.

The paper introduces NuFold, a tool for large-scale RNA tertiary structure prediction, and provides a database with precomputed structures and confidence metrics.

Distance-AF Multimer improves protein complex structure prediction by incorporating user-defined residue distance constraints, achieving competitive or better performance than existing methods.

DAQplugin provides real-time deep learning-based validation of cryo-EM protein models in ChimeraX, identifying residue errors and offering correction guidance.

PathFold predicts protein folding pathways from sequence using a deep learning framework combining AlphaFold and diffusion models, showing agreement with experimental data.

Prot-LAMBDA improves protein structure prediction by explicitly incorporating spatial relationships, achieving better performance than ESM2-3B in contact prediction and 3D structure prediction.

Queryome is a deep research system that improves biomedical literature synthesis through iterative reasoning and retrieval, achieving higher accuracy than existing methods.

The paper validates a cell-free nucleosome assembly method using ExACT and confirms its structural fidelity through cryo-EM and deep-learning-based model quality scoring.

DMcloud improves fitting of molecular models into cryo-EM maps by local structure fitting, achieving higher accuracy and coverage than existing methods.

This paper reports on the fourth CAFA challenge, showing improvements in Gene Ontology term prediction but highlighting challenges in new evaluation settings.

Emap2lig detects ligand densities and builds atomic structures from cryo-EM maps using a two-stage deep learning framework.

PL-PatchSurfer3 improves structure-based virtual screening by incorporating hydrogen bond complementarity and visibility features, enhancing performance while maintaining tolerance to receptor structure changes.

CryoZeta improves cryo-EM structure modeling by integrating density features with sequence data using a diffusion-based deep learning model, achieving higher accuracy than existing methods.

The paper investigates the structural differences between two capsid sizes in bacteriophage P1, revealing the role of protein appendages in DNA packaging and capsid morphogenesis.

This paper evaluates blind predictions of dynamic water and ion ensembles around RNA using cryo-EM densities from CASP16, finding that molecular dynamics approaches perform best.

This study identifies proteomic changes in Crocosphaera subtropica during light-dark cycles and nitrogen fixation, revealing coordinated regulation of metabolic pathways and nitrogenase expression.

Distance-AF improves AlphaFold2 predictions by incorporating user-specified distance constraints to refine protein structures and reduce RMSD.

CryoDataBot automates the creation of high-quality cryoEM datasets for AI-driven structural biology, improving training efficiency and enabling flexible dataset customization.

This paper improves AlphaFold's self-confidence scoring by modifying its LDDT prediction head to provide more accurate model quality assessments.

The paper uses cryo-EM to determine the atomic structure of the P1 Tail Baseplate complex, revealing how its unique infection mechanism involves LTFs and baseplate movement.

DistPepFold improves protein-peptide complex docking using a privileged knowledge distillation approach based on AlphaFold-Multimer, outperforming AFM on two datasets.

This study uses proteomic analysis to examine how extended light or dark conditions affect protein expression in the cyanobacterium Crocosphaera subtropica, identifying key proteases involved in photosynthetic regulation.

NuFold predicts RNA tertiary structures using deep learning with a nucleobase center representation to bridge the gap between sequence data and structural insights.

The paper predicts transcriptional activation domain function using graph neural networks, achieving high accuracy and outperforming existing methods.

GO2Sum generates human-readable protein function summaries from GO terms using a fine-tuned T5 model, outperforming the base T5 model in creating function descriptions for UniProt entries.

The paper introduces Mem-LZerD, a membrane protein docking algorithm that improves the modeling of membrane protein complexes using geometric hashing and membrane constraints, achieving higher success rates than previous methods.

DiffModeler uses diffusion models to accurately model protein complex structures from low-resolution cryo-EM maps, achieving high TM-Scores across various resolutions.

AFM-RL predicts structures of large protein complexes with six to twenty chains using AlphaFold-Multimer and reinforcement learning, showing improved performance over existing methods.

The paper introduces a method combining deep learning for density tracing with AlphaFold2 to improve protein structure modeling from cryo-EM maps and accurately assign chain identities in homo-multimers.

The paper introduces Domain-PFP, a method for protein function prediction using function-aware domain embeddings, which outperforms existing methods and large-scale protein language models in GO prediction tasks.