graph neural network

This paper develops a graph-based framework for label-free detection of nanopore translocation events, showing it can predict event morphology across a range of voltages.

The paper introduces a quantum-walk-based ranking method to stabilize neoantigen selection in near-tie scenarios by leveraging graph structure and coherence dynamics.

This paper develops a multi-modal deep learning framework to identify Class I HDAC inhibitors as broad-spectrum anti-cancer agents by integrating chemical and transcriptomic data.

The paper develops GENIA, a machine learning framework for designing synthetic microbial communities that degrade multiple persistent pollutants more effectively than existing methods.

SpaMosaic integrates spatial multi-omics data using contrastive learning and graph neural networks, improving spatial domain identification and missing modality imputation across diverse datasets.

This paper explores the use of Graph Neural Networks for predicting protein-ligand interactions, focusing on improving interpretability and accuracy through various techniques.

SEAGALL uses geometry-aware graph attention networks to extract biologically relevant features from single-cell data, revealing cell identity drivers beyond marker genes.

EpiRanha identifies epitope similarities using a structure- and residue-aware graph neural network, improving detection of cross-reactive epitopes and off-target interactions.

This paper introduces a proximity-aware graph attention network to infer cell-cell communication from spatial transcriptomics data, achieving improved performance over existing methods.

This paper introduces a hierarchical spatial graph neural network to classify tertiary lymphoid structures as immunogenic or tolerogenic in renal cell carcinoma using spatial transcriptomics data.

Concordia detects spatial domains in population-level spatial proteomics data using a GNN-based framework, revealing a cancer-associated fibroblast subset linked to clinical outcomes in lung cancer.

This paper introduces a graph transformer model for drug-target binding affinity prediction, showing improved performance on benchmark datasets.

This study benchmarks kinase bioactivity models across splitting strategies and protein representations, finding that Random Forest with Morgan fingerprints outperforms neural models in scaffold and target-based evaluations.

The paper develops a structure-informed Siamese GNN to classify CirA missense variants affecting cefiderocol susceptibility using synthetic data.

The paper uses FAMPNN to redesign SrBDH1 for increased thermostability while maintaining stereoselectivity.

The paper introduces SCRU-DB, a database of RNA units, and two models, SCRU-Seq and SCRU-Diff, to design RNA sequences with high structural accuracy.

GraphPop reduces population genomics complexity by using graph databases, enabling faster analysis of large datasets and revealing population structure in rice and humans.

GraphMana manages population genomics data using a graph database, enabling efficient incremental updates and export to multiple formats.

This paper introduces GNN-GBA, a graph neural network that identifies mechanistic pathways linking gut microbes to brain disorders using a large biomedical knowledge graph.

This paper introduces IARA, a deep learning model that rapidly identifies optimal binding interfaces for protein binder design, reducing computational costs.

KIMMDY emulates biomolecular reactions using kinetic Monte Carlo and graph neural networks, validated against experimental data and applied to various reaction types.

VASCIF predicts antibody-antigen interfaces using a structure-aware graph model, achieving state-of-the-art performance and revealing biophysically meaningful interaction patterns.

DyGraphTrans models disease progression from EHRs using a dynamic graph framework, achieving strong predictive performance and interpretability.

OmniCleave models protease-substrate specificity using a geometric graph framework, outperforming existing methods and identifying novel substrates.

This paper develops a multi-stage graph attention network to improve Alzheimer's disease classification from genome-wide association data by capturing epistatic interactions and enhancing interpretability.

The paper introduces MoCL-GAT, a self-supervised learning framework for molecular representation, achieving state-of-the-art performance on molecular property prediction tasks.

ARGMix uses a graph transformer to infer local ancestry from ancient DNA, showing continuity between Otzi the Iceman and present-day Europeans.

XAttn-DTA predicts drug-target affinity using graph attention networks and cross-attention fusion, achieving improved performance on benchmark datasets.

CLIMBS is a graph neural network classifier that assesses carbohydrate-protein interactions using synthetic negative data, showing high accuracy and specificity.

GATSBI improves protein embeddings by using biologically relevant data splits, showing better performance for understudied proteins in various tasks.