graph neural network

This paper introduces GraphEBM, an SE(3)-equivariant graph attention network for protein side chain conformation, outperforming existing methods in rotamer recovery and combinatorial optimization.

HERO improves error correction in long reads by combining de Bruijn and overlap graphs, reducing indel and mismatch errors by 65% and 20% respectively.

This paper introduces a graph matching method to track neurons in freely-moving C. elegans, achieving over 75% accuracy with five labeled frames.

Pasa improves genome scaffolding by using population pangenome graphs, enhancing accuracy and efficiency for bacterial genome assemblies.

LightMHC predicts full-atom pMHC structures from sequences using a lightweight model with attention mechanisms, achieving comparable performance to AlphaFold 2 and ESMFold with faster speed.

LightRoseTTA provides a high-efficient and accurate protein structure prediction method, outperforming RoseTTAFold in speed and parameter count while maintaining competitive accuracy.

This paper introduces TCGN, a deep learning model that integrates CNN, GNN, and transformer for gene expression prediction from histopathological images, achieving higher accuracy and interpretability.

The paper introduces graph neural network models to identify moonlighting proteins, achieving high accuracy on human data and offering a lightweight approach for larger datasets.

CGCom uses a graph neural network to infer cell-cell communication from spatial transcriptomics data, showing improved accuracy over existing methods.

This paper introduces REGNN, a graph deep learning framework for analyzing spatially resolved transcriptomics data, showing improved performance in identifying mosaic-like tissue architectures in kidney diseases.

Speos predicts core genes for complex diseases using graph representation learning, achieving validation rates comparable to Mendelian disorder genes and identifying druggable targets.

This paper introduces BrainGAT, a graph neural network for gender classification using multimodal brain connectomes, achieving 83.13% accuracy on the Human Connectome Project dataset.

The paper introduces GraCoal embedding to improve functional organization analysis of scale-free biological networks by incorporating graphlet patterns, outperforming traditional Spring embedding methods.

CONE generates context-specific gene embeddings from context-free interactomes to identify disease-associated genes using contextualized random walks and graph attention networks.

This study develops AOMP, a graph neural network for predicting AOX metabolism, outperforming benchmarks and identifying new scaffolds validated experimentally.

This paper introduces GRAPE-Net, a graph-based model that stratifies bronchial premalignant lesions using histopathology images, showing improved classification accuracy for lung cancer subtypes.

This study introduces BINARY, a graph convolutional network using binarized spatial transcriptomics data to identify spatial domains, outperforming existing methods with less input.

This paper introduces a general single-cell analysis framework using conditional diffusion models, showing competitive performance across tasks with added flexibility via prior information.

This paper develops STAligner, a graph attention neural network for integrating spatial transcriptomics data across different conditions, technologies, and developmental stages, identifying shared and specific spatial domains.

This paper introduces Net-DMPred, a graph neural network-based method to predict cancer-specific driver missense mutations by incorporating molecular network structures.

This paper introduces a method to transform EEG data into a spectral representation using learned graphs for motor imagery decoding, showing improved performance over existing methods.

This work introduces an equivariant graph neural network to predict tertiary structural changes from single-site mutations, outperforming AlphaFold in mutant structure prediction.

This study shows that CRF1 antagonist R121919 and naltrexone restore prefrontal cortex functional connectivity during alcohol withdrawal through distinct mechanisms involving modularity and connectivity changes.

Nanotope predicts nanobody paratopes using graph neural networks and sequence/structural features, achieving higher accuracy than traditional methods without requiring antigen data.

GLMSite identifies DNA and RNA-binding sites using geometric graph learning on ESMFold-predicted structures, outperforming sequence-based methods and matching structure-based approaches.

GraphPart partitions biological sequences to avoid closely related sequences in different partitions, retaining more sequences than reduction methods while maintaining homology separation quality.

This paper introduces PGCN, a structure-aware graph convolutional network for predicting protease specificity and designing proteases, showing accurate predictions and successful experimental validation.

This paper introduces EGGN, an exemplar guided graph neural network for predicting gene expression from spatial transcriptomics tissue slide images, showing superior performance on benchmark datasets.

The paper evaluates the gRNAde software on updated RNA structure data, confirming its ability to predict sequences from structures but noting lower performance on ligand-bound RNAs like riboswitches.