graph neural network

STEP: Spatial Transcriptomics Embedding Procedure for Multi-scale Biological Heterogeneities Revelation in Multiple Samples

STEP is a novel AI architecture for spatial transcriptomics data that improves the identification of cell types and spatial domains, with enhanced accuracy in liver zone demarcation.

A graphSAGE discovers synergistic combinations of Gefitinib, paclitaxel, and Icotinib for Lung adenocarcinoma management by targeting human genes and proteins: the RAIN protocol

The paper proposes a method using graph neural networks to recommend drug combinations for lung adenocarcinoma by targeting genes and proteins, finding Gefitinib, Paclitaxel, and Icotinib as the most effective combination.

Identifying new cancer genes based on the integration of annotated gene sets via hypergraph neural networks

This study introduces DISHyper, a hypergraph neural network that integrates annotated gene sets to predict cancer genes, identifying 44 novel ones associated with multiple cancers.

Spatial motifs reveal patterns in cellular architecture of complex tissues

The paper introduces a method to detect spatial motifs in tissue maps, revealing previously unrecognized patterns in cell type arrangements in mouse neural tissues.

scGHOST: Identifying single-cell 3D genome subcompartments

SCGHOST identifies single-cell 3D genome subcompartments using scHi-C data and provides insights into cell-to-cell variability in nuclear organization.

Towards improving full-length ribosome density prediction by bridging sequence and graph-based representations

This study introduces RiboGL, a model combining graph and recurrent neural networks to predict ribosome density from mRNA secondary structure, outperforming existing methods.

gRNAde: Geometric Deep Learning for 3D RNA inverse design

gRNAde designs RNA sequences based on 3D backbone structures using geometric deep learning, achieving higher native sequence recovery than Rosetta on benchmark tasks.

Labels as a Feature: Network Homophily for Systematically Discovering human GPCR Drug-Target Interactions

The paper introduces a model called CSNN that uses network homophily to improve in-distribution predictions of drug-target interactions for human GPCRs.

MegaKG: Toward an explainable knowledge graph for early drug development

This paper constructs a comprehensive biomedical knowledge graph, MegaKG, integrating 23 data sources to support drug development by generating explainable hypotheses and facilitating practical applications.

GraphMHC: neoantigen prediction model applying the graph neural network to molecular structure

GraphMHC predicts neoantigen binding using graph neural networks on molecular structures, achieving high accuracy and showing survival differences in melanoma patients.

ProInterVal: Validation of Protein-Protein Interfaces through Learned Interface Representations

This paper introduces a method to validate protein-protein interfaces using learned representations, achieving 0.91 accuracy on a benchmark dataset.

Water position prediction with SE(3)-Graph Neural Network

The paper introduces WatGNN, a low-computation water position prediction method using SE(3)-equivariant graph neural networks, achieving state-of-the-art performance in 1.86 seconds on average.

Connectome spectrum electromagnetic tomography: a method to reconstruct electrical brain source-networks at high-spatial resolution

This paper introduces CSET, a method that reconstructs electrical brain source-networks with high spatial resolution using diffusion MRI and EEG data.

AOPWIKI-EXPLORER: An Interactive Graph-based Query Engine leveraging Large Language Models

The paper introduces an interactive graph-based query engine for AOP wiki data using large language models to enhance exploration and integration of adverse outcome pathway networks.

HyGAnno: Hybrid graph neural network-based cell type annotation for single-cell ATAC sequencing data

HyGAnno transfers cell type labels from scRNA-seq to scATAC-seq using a graph neural network, improving annotation accuracy and robustness.

PROTACable is an Integrative Computational Pipeline of 3-D Modeling and Deep Learning to Automate the De Novo Design of PROTACs

PROTACable automates PROTAC design using 3-D modeling and deep learning to predict activity of newly designed PROTACs.

SPDesign: protein sequence designer based on structural sequence profile using ultrafast shape recognition

SPDesign designs protein sequences using structural sequence profiles and ultrafast shape recognition, outperforming existing methods in sequence recovery and structural accuracy.

Simultaneous Inference of Past Demography and Selection from the Ancestral Recombination Graph under the Beta Coalescent

This paper introduces two methods to infer demographic history and selection from sequence data by modeling the ancestral recombination graph under the beta coalescent.

Graphical pangenomics-enabled characterisation of structural variant impact on gene expression in Brassica napus

This study uses pangenome graphs to analyze structural variants in Brassica napus, showing they impact gene expression and eQTL analysis better than single references.

Detection of spatial chromatin accessibility patterns with inter-cellular correlations

This paper introduces Descartes, a model for identifying spatially variable chromatin accessibility peaks using inter-cellular correlations, showing improved performance and efficiency over existing methods.

GATLncLoc+C&S: Prediction of LncRNA subcellular localization based on corrective graph attention network

The paper introduces GATLncLoc+C&S, a model for predicting LncRNA subcellular localization using a corrective graph attention network, achieving high accuracy on benchmark datasets.

Integration of graph neural networks and genome-scale metabolic models for predicting gene essentiality

The paper introduces FlowGAT, a hybrid FBA-machine learning method to predict gene essentiality in E. coli by leveraging metabolic network structure and knock-out fitness data.

Fast protein structure searching using structure graph embeddings

The paper introduces Progres, a method for fast protein structure searching using graph neural networks, achieving accuracy comparable to existing methods with rapid query times.

Graph Structured Neural Networks for Perturbation Biology

This paper introduces Graph Structured Neural Networks (GSNN) for perturbation biology, showing they outperform baselines in predicting perturbed gene expression, drug combinations, and disease-specific drug prioritization.

Multi-view graph learning for deciphering the dominant cell communication assembly of downstream functional events from single-cell RNA-seq data

This paper introduces scDCA, a method to identify dominant cell communication assemblies from scRNA-seq data, validated in renal cell carcinoma to reveal key immune gene expression patterns and therapeutic responses.

Modeling cancer dependency with deep graph models

The paper introduces a deep learning framework called DepGPS to model tumor vulnerabilities using graph neural networks and transformers, showing improved prediction of context-specific vulnerabilities and utility in identifying synthetic lethal genes.

Automated model building and protein identification in cryo-EM maps

ModelAngelo combines convolutional and graph neural networks with protein language model embeddings and hidden Markov model searches to automate atomic model building and identify unknown proteins in cryo-EM maps.

Inferring Dynamic Functional Connectivity from Field Potentials Using Graph Diffusion Autoregression

This paper introduces a graph diffusion autoregressive model to estimate dynamic functional connectivity from neural field potentials, validated on simulated data and macaque recordings.

MaskGraphene: an advanced framework for interpretable joint representation for multi-slice, multi-condition spatial transcriptomics

MaskGraphene integrates multi-slice, multi-condition spatial transcriptomics data using a graph neural network to generate interpretable joint embeddings with preserved geometric information, improving downstream analyses.

Quantitative characterization of cell niches in spatial atlases

The paper introduces NicheCompass, a graph deep learning method for characterizing cell niches in spatial omics data, showing improved performance in identifying and analyzing niches across multiple biological contexts.