graph neural network

This study uses network analysis in Drosophila melanogaster to identify mitochondrial genes associated with aging and age-related diseases.

EquiScore is a protein-ligand interaction scoring method that outperforms existing methods on external test sets and shows potential for lead compound optimization.

PLANET is a graph neural network model for protein-ligand binding affinity prediction that outperforms existing methods in virtual screening while being computationally efficient.

This paper introduces DeepBindGCN, a model that predicts protein-ligand interactions using graph convolutional networks and molecular vectors, achieving high accuracy in benchmark datasets.

This paper introduces a method using fMRI spectral signatures on high-resolution cortical graphs to improve individual identification by addressing limitations of conventional connectivity-based approaches.

The paper proposes a data processing pipeline that generates graph representations of iEEG data and evaluates their effectiveness for seizure detection using graph neural networks.

This paper introduces a GCN-based method for automating and interpreting scRNA-seq cell type annotation, showing improved accuracy over consensus and neighbor majority approaches.

This paper improves drug repurposing by integrating 2D drug molecular structures into graph neural networks, achieving higher AUROC and AUPRC scores than previous methods.

This paper introduces GraphGPSM, a graph neural network-based scoring model for protein structure prediction, which improves accuracy over existing methods on benchmark datasets.

SIMBA is a graph embedding method that jointly embeds single cells and their features into a common space, enabling diverse single-cell analysis tasks.

This paper introduces DProQA, a gated graph transformer for assessing the quality of predicted protein complex structures, achieving third place in CASP15 quality assessment.

The paper uses a graph database to identify fragment merges from chemical catalogues, finding complementary results to traditional similarity searches.

This paper introduces TransFun, a method that combines protein sequences and structures using transformers and equivariant graph neural networks to improve protein function prediction.

This paper introduces GraNA, a deep learning framework for supervised biological network alignment that improves functional annotation transfer between species.

This paper introduces GNN-C2L, a spatial cell-type deconvolution method that improves performance by incorporating spatial inductive biases, outperforming spatial-agnostic variants on simulated datasets.

Stability Oracle predicts stabilizing mutations using a structure-based graph-transformer, achieving high accuracy on thermodynamic stability changes.

This study evaluates regulatory network-based machine learning for gene expression prediction in maize using Arabidopsis thaliana networks, finding that simple models outperform complex ones within species but complex models are better for cross-species predictions.

ZymePackNet predicts protein sidechain conformations using a graph neural network without requiring rotamer libraries or energy functions, achieving high accuracy in dihedral angle prediction.

This paper introduces MoCoP, a contrastive pretraining framework that combines molecular graphs and cellular morphologies to improve QSAR model performance.

This study introduces a graph-based method to infer temporal signaling networks from high-resolution phosphoproteomics data, improving the identification of signaling-relevant interactions.

SpatialGlue integrates spatial multi-omics data using a graph neural network to improve spatial domain detection and cross-modality correlation analysis.

The paper introduces voxel-wise brain graphs from diffusion MRI, showing that functional MRI data can be approximated by low-frequency graph Laplacian eigenmodes.

This paper introduces iMFP-LG, a method combining protein language models and graph attention networks to identify multi-functional peptides, identifying 8 candidates and confirming one with antibacterial and anticancer effects.

The paper introduces a large-scale spatial transcriptomic dataset and a Graph Neural Network-based method for cell type deconvolution in human immuno-oncology.

SLAT aligns heterogeneous spatial omics data by integrating omics profiles and spatial context, improving cell-typing resolution and regulatory inference.

The paper compares sequence-based and structure-based methods for predicting protein function, finding that structure-based approaches can offer better accuracy when using high-quality protein structures.

MOGAT improves multi-omics integration for cancer subtype prediction by addressing limitations in existing graph neural network approaches.

GraphBepi predicts B-cell epitopes using AlphaFold2 structures and a pretrained language model, outperforming existing methods in AUC and AUPR.

SUPREME predicts cancer subtypes using multiomics data and graph convolutions, outperforming existing methods and showing significant survival differences.

This paper introduces an attention-based GCN method for breast cancer subtype classification using multi-omics data and identifies patient-specific gene markers through interpretability techniques.