graph neural network

StressNET: an adaptable deep-learning model for mechanical stress inference in tissues

StressNET infers intercellular mechanical stresses from tissue geometry using a Graph Neural Network, achieving state-of-the-art performance in zebrafish and Xenopus embryos.

Phylogenies as graphs: structured neural networks improve host origin predictions from paramyxovirus sequences

This paper uses graph neural networks to predict host origins from paramyxovirus sequences, showing improved performance over traditional methods, especially with sparse data.

POAnoise: A Graph-based Denoising Pipeline for Amplicon Sequencing Data

POAnoise improves amplicon sequencing denoising by using graph-based alignment and abundance-aware clustering, achieving higher F1-scores and more stable performance than existing methods.

Somatic haplotype reconstruction and variant recalibration from tumor-only long-read sequencing

LongPhase-TO reconstructs somatic haplotypes from tumor-only long-read sequencing, improving variant calling accuracy and haplotype block size.

Pep-PU-GAN: Positive-Unlabeled Adversarial Learning for Peptide Function Prediction

Pep-PU-GAN improves peptide function prediction using PU learning and GANs with graph neural networks, achieving high F1 and AUROC scores on neuropeptide classification.

Gaze Graphs in spatial navigation: visual behavior during free exploration explains individual differences in navigation performance

This study uses gaze graphs from eye-tracking data during virtual reality exploration to identify predictors of individual differences in spatial navigation performance.

Structured cross-omics interaction discovery with a triple-graph model

The paper introduces TriGer, a triple-graph model that identifies cross-omics modules by combining cross-layer associations with dependency structures, showing effectiveness in inflammatory bowel disease and colorectal cancer.

POME: Graph-based embeddings for partially observed mixed-type data

POME provides graph-based embeddings for partially observed mixed-type biomedical data, achieving state-of-the-art imputation and enabling downstream analyses like subgroup discovery and predictive modeling.

Heterogeneous graph neural networks with biological prior knowledge for interpretable drug repurposing in triple-negative breast cancer

PRECISION uses a heterogeneous graph neural network with biological knowledge to identify drug repurposing candidates in TNBC, showing associations between predicted drug sensitivity and survival in multiple cohorts.

Resolving context-specific protein-protein interactomes forbiological discovery and therapeutic target prioritisation

ProtScape learns context-specific protein interactomes across 200 cell types, improving interaction reconstruction and identifying disease-related mechanisms.

TLS-Sim: An Open-Source Toolkit for Individualized Transcranial Light Stimulation Modeling with Deep-Learning-Accelerated Simulation

TLS-Sim provides a toolkit for individualized transcranial light stimulation modeling using deep-learning-accelerated Monte Carlo simulations to improve efficiency and accuracy.

Himito: a Graph-based Toolkit for Mitochondrial Genome Analysis using Long Reads

Himito is a graph-based toolkit for analyzing mitochondrial genomes using long reads, achieving superior performance in assembly and variant calling and revealing age-related patterns in the All of Us dataset.

A heterogeneous biomedical knowledge network framework for rare disease drug candidate prioritization: integrating Orphadata and DisGeNET via gene-bridge harmonization

This paper develops a network framework to prioritize drug candidates for rare diseases by integrating Orphadata and DisGeNET through gene-bridge harmonization and using a Graph Attention Network for node representation learning.

ProMaya: a hierarchical universal Deep Learning framework for accurate and interpretable Protein-Protein interaction identification

ProMaya identifies protein-protein interactions using a hierarchical graph-transformer framework, achieving >95% accuracy across nine species and outperforming existing tools by >12%.

Generalizable Cysteine Quantification in Pea Cultivars from SERS Spectra Using AI

The paper uses AI to quantify cysteine in pea cultivars from SERS spectra, finding that a 1D-CNN outperforms other models in accuracy and generalization.

A network-based framework for detecting communities of similar neural spike trains

This paper introduces a network-based method to detect neuronal assemblies from spike train data using the van Rossum distance and Louvain algorithm, validated on synthetic and real datasets.

Benchmark validity in graph neural network scoring of metabolic reaction activity on Recon3D: detecting label leakage, memorized noise and input-invariant models

The paper benchmarks a graph neural network for scoring metabolic reaction activity in tumors but finds issues like label leakage and memorization, leading to retraction of previous results.

GIN-CRC-Pareto: A graph-based Pareto-optimal multi-task learning framework to identify miRNA-target interactions in colorectal cancer

GIN-CRC-Pareto identifies miRNA-target interactions in colorectal cancer using a graph-based multi-task learning framework, outperforming existing methods in accuracy and AUC.

Relational Graph Convolutional Networks for Glioblastoma Biomarker Discovery via ceRNA and Copy Number Variation Analysis

This paper uses relational graph convolutional networks to integrate ceRNA and copy number variation data for glioblastoma biomarker discovery, identifying five novel candidate biomarkers.

Pangenome alignment reveals global diversity and evolution of human centromeric regions

The paper introduces Centrolign, a tool for aligning centromeric regions, revealing mutation dynamics and enabling variant calling in human centromeres.

BioGraphX-RNA: A Universal Physicochemical Graph Encoding for Interpretable RNA Subcellular Localization Prediction

BioGraphX-RNA predicts RNA subcellular localization using physicochemical graph encoding, showing competitive performance and interpretability across mRNA, miRNA, and lncRNA.

Atlantis: An integrative database for human proteome structural and functional sites

Atlantis integrates structural and functional residue-level data for the human proteome, enabling complex queries and analysis of protein interactions and mutations.

Comprehensive Evaluation of Protein Language Model Embeddings for Drug-Target Affinity Prediction

This study evaluates the impact of pre-trained protein language models on drug-target affinity prediction, showing that simple architectural changes can match large PLMs' performance.

MHCXGraph: A Graph-Based approach to detecting T cell receptor cross-reactivity

MHCXGraph identifies conserved structural regions in pMHC complexes to detect T cell cross-reactivity using graph-based analysis.

Cross-tissue Graph Attention Networks for Semi-supervised Gene Expression Prediction

This paper introduces gemGAT, a graph attention network for gene expression imputation across tissues, showing improved performance over regression-based methods on GTEx and ADNI data.

TomatoPGFM: A graph-conditioned foundation model for tomato pangenomes

TomatoPGFM is a graph-conditioned foundation model for tomato pangenomes that improves sequence representation using graph context, achieving higher classification performance than DNABERT-2.

Pro-GAT: Reconnecting Fragmented PROTACs Using Graph Attention Transformer

Pro-GAT repairs chemically disconnected PROTAC candidates generated by diffusion models, improving validity while maintaining uniqueness.

Lineage-specific X chromosome inactivation escape and skew underlie sex-biased immune gene dosage and deleterious variant exposure

This paper introduces scDaisyChain to analyze X chromosome inactivation escape and skew in immune cells, revealing lineage-specific effects on gene dosage and variant exposure in healthy and autoimmune disease contexts.

UMITIC: An unsupervised framework for the joint characterization of cellular phenotypes and spatial neighborhoods in multiplex and hyperplex immunofluorescence imaging data

UMITIC is an unsupervised framework that characterizes cellular phenotypes and spatial neighborhoods in multiplex imaging data, showing improved accuracy over existing methods.

Benchmarking Graph Neural Networks for Multi-Omics Cancer Subtyping using Methylation and Gene Expression Profiles

This study benchmarks graph neural networks for cancer subtyping using methylation and gene expression data, finding that ChebNet performs best and simple readout layers improve performance.