graph neural network

Somatic structural variation discovery on cancer pangenome graphs

JANUS discovers somatic structural variants in cancer genomes using a pangenome graph approach, achieving high accuracy and efficiency compared to traditional methods.

Classifying tumors from broad category to hierarchical molecular subtype: an interpretable hierarchical framework for precision oncology

This paper introduces HHRM, a hierarchical model for tumor subtype classification, achieving high accuracy in pan-cancer and breast cancer subtyping with interpretable biomarker identification.

Deep Recovery of Effective Migration Surfaces

This paper introduces dreems, a graph neural network that efficiently estimates effective migration surfaces from genetic data, achieving lower error and faster computation than existing methods.

Graph neural network approach for inferring neuronal connectivity

This paper introduces CircuitGNN, a graph neural network that infers neural connectivity from population activity, showing high accuracy and robustness across different circuit sizes and recording conditions.

FateDriver: Weakly supervised dual-view graph learning for transition-specific prioritization of candidate regulators in cellular reprogramming

FateDriver prioritizes candidate regulators for specific cellular transitions using a weakly supervised dual-view graph neural network, outperforming existing methods in recovering known regulators.

Interpreting graph neural networks reveals microbial association patterns in inflammatory bowel diseases

The paper uses graph neural networks to interpret microbial association patterns in IBD, showing that coordinated associations, not just abundance shifts, are predictive.

Graph-based pangenome analysis uncovers structural and functional impacts of allopolyploidization events

The paper analyzes allopolyploidization events in Brettanomyces bruxellensis using graph pangenomes to identify structural variants and their functional impacts.

Interpreting Graph Neural Networks for Antibacterial Drug Repurposing against Acinetobacter baumannii

This paper uses a graph neural network to predict antibacterial activity against Acinetobacter baumannii and identifies piperazine as a key contributor to predicted activity, leading to the testing of CE3F4.

De novo Protein Generating Model via Structure-Aware Graph Neural Networks

This paper introduces a protein conformation prediction model optimized for structural accuracy using a structure-aware GNN, achieving results comparable to sequence-focused methods without sequence recovery supervision.

Multiplex-GCN Predicts Clinically Actionable Tumor Microenvironment Subtypes in Pancreatic Cancer

The paper develops a multi-layered GCN to identify TME-guided PDAC subtypes, revealing seven subtypes with distinct molecular and prognostic features.

Comparative analysis identifies a shared topological organisation among divergent intergenic RNAs

The paper identifies a shared topological organisation in divergent arenavirus intergenic RNAs using graph theory and structural analysis.

Core Coefficient quantifies transcription-factor functional dominance for therapeutic target selection

The paper introduces the Core Coefficient (CC) to quantify transcription factor functional dominance in regulatory networks, showing it can predict therapeutic target shifts in disease progression.

Predicting clonal mutations from tumour spatial transcriptomics data using a graph convolutional neural network

This paper introduces SPOT-Mut, a graph convolutional neural network that predicts clonal mutations from spatial transcriptomics data, achieving high precision and recall for key driver genes.

An Interactive Bidirectional De Novo Protein Design System

This paper introduces an interactive server that combines a structure-aware GNN and ESMFold for bidirectional protein design, enabling users to analyze and validate design hypotheses without software installation.

A rare pre-existing progenitor-like Primed SMC compartment is the dominant inferred source of SMC-derived cellularity in vascular injury and atherosclerosis

Overcoming segmentation bottlenecks in brain tissue for spatial proteomic inference at population scale

The paper introduces SPInfer, a workflow for spatial proteomic inference that improves cell segmentation in human brain tissue, enabling better biological discovery at population scale.

Phenotype-cloud reorganization shapes cancer progression and therapeutic adaptation

This study identifies phenotype-cloud transitions as an organizational feature of cancer progression and therapeutic adaptation, showing how tumours reorganize phenotypic states during progression and treatment.

geoGraph: Connectivity modelling across geographic space with hexagonal discrete global grids

geoGraph models organism movement and connectivity across geographic space using hexagonal grids, improving accuracy over rectangular grids by avoiding distortion.

β-Stability: Efficient, human-in-the-loop, graph-based curation of individual animal identities from images

The paper introduces β-Stability, a graph-based framework for semi-automated curation of animal identities from images, achieving high accuracy with limited expert review.

Relation-Aware Graph-Hypergraph Networks for Antibody Sequence-Structure Co-Design

This paper introduces RA-GHN, a framework for antibody sequence-structure co-design that captures multi-body interactions and structural constraints, outperforming existing methods.

qcGEM: a graph-based molecular representation with quantum chemistry awareness

qcGEM introduces a quantum-chemistry-aware graph-based molecular representation that improves performance and interpretability in molecule-related tasks.

LigRNA: High-accuracy prediction of RNA-ligand binding sites via dynamic chemical complementarity and language model encoding

LigRNA predicts RNA-ligand binding sites using a graph neural network that combines language model embeddings and physicochemical features, achieving improved performance over existing methods.

NoiTox: A Machine Learning Strategy for Peptide Toxicity Prediction Reveals Label Noise in Benchmark Datasets

NoiTox presents a multi-model approach for peptide toxicity prediction, revealing label noise in datasets and showing that physicochemical descriptors are informative.

Edge-Aware Graph Attention Networks for Interpreting Biophysical Mechanisms from Molecular Dynamics Simulations

This paper introduces edge-aware graph attention networks to interpret biophysical mechanisms from molecular dynamics simulations, showing improved prediction and communication network analysis.

Learning Dynamic Protein Representations at Scale with Distograms

Adding Boltz2 distogram edges and features to protein and RNA graph networks beats MD-correlation graphs on binding tasks and improves stability, interaction and RNA predictions without simulation.

OmniScore: Universal Scoring of Diverse Biomolecular Complexes via Equivariant Geometry-Aware Discrete Representation Learning

OmniScore learns shared geometric representations of diverse biomolecular complexes and improves quality assessment and interaction prediction across molecular classes.

Histology-aware Inference of Cell-cell Communication in Spatial Transcriptomics

This paper introduces HARMONIC, a framework that improves cell-cell communication inference in spatial transcriptomics by incorporating histology data, showing it better captures anatomical context.

Filling the Missing Puzzle: Integrating Interaction and Evolutionary Information to Enhance Protein Fitness Prediction in Inverse Folding

This paper introduces zPolaron, a graph neural network for functional protein design that integrates structural, interaction, and evolutionary information, improving fitness prediction beyond structural compatibility.

AI-Driven Design of Next-Generation Immunoinformatics Multi-Epitope Subunit Vaccine Targeting the Most Virulent Mpox Proteome: A Structural and Kinetic Validation Framework

This paper develops an AI-driven vaccine design framework for Mpox, yielding three multi-epitope constructs validated for structural stability, immune binding, and expression efficiency.

IsoGraph resolves coordinated transcript choice and gene abundance in the aging human brain

IsoGraph identifies age-associated transcript-usage signals in human brain transcriptomes, distinguishing them from gene abundance and cellular composition effects.