graph neural networks

Bond-Level Prediction of U-Donor Bond Lengths in Uranyl Coordination Complexes Using Machine Learning for Ligand Recognition

Abstract U-donor bond lengths in uranyl coordination complexes directly reflect metal–ligand interaction strength and provide important structural parameters for evaluating ligand recognition, selective complexation, and extraction/adsorption system design. To enable bond-level prediction, this study developed a DFT-de...

Core code and synthetic examples for characterizing three-dimensional urban canopy structure in Shanghai

This record provides author-developed core code and synthetic examples accompanying the manuscript “Characterizing three-dimensional urban canopy structure in Shanghai with spatial graph learning”. The package includes implementations for attribute-space and spatially regularized K-means, graph neural network encoders,...

Harnessing artificial intelligence for fleet decarbonization: a Tunisian case study using deep reinforcement learning and graph neural networks

NORMAN aquatic ecotoxicity model v0.1: QMRF, QPRF example, model card and SusDat predictions

A multi-species model for the acute and chronic aquatic toxicity of organic chemicals, developed for the NORMAN network. It predicts toxicity values (log10 µg/L) for 121 species-and-endpoint combinations (heads), of which 82 are released for use (59 acute: LC50/EC50; 23 chronic: NOEC/EC10), covering fish, crustaceans,...

From Faraday and Maxwell to Neural Cycle Topology: The Operations of a Coupled Conductive-Loop Lattice

We derive the governing equation of neural cycle topology (NCT) computing from Maxwell’s equations and set what follows from it against measurements of a 362-cavity lattice read by a pickup on every cavity. In the magnetoquasistatic limit, charge conservation confines the conductor currents to the cycle space of the ne...

Graph-Based Analysis of Functional Brain Connectivity from fMRI Data

Functional magnetic resonance imaging (fMRI) provides whole-brain measurements that can be converted into functional connectivity networks for quantitative analysis. Graph-based methods represent brain regions as nodes and statistical relationships among regional blood oxygen level-dependent signals as edges in this re...

A relation-aware heterogeneous graph neural network for the hierarchical regionalization of urban morphology

Collagen EDC/NHS–gallic Vinardo docking tables, receptors, and HeteroGNN LOLO metrics

Supporting dataset for a nine-ligand EDC/NHS–gallic acid type I collagen crosslinking study that maps non-covalent binding with SMINA/Vinardo docking and evaluates a post-hoc HeteroGNN surrogate under leave-one-ligand-out (LOLO) hold-out. Contents. (i) Collagen docking score tables (~6,156 poses across 76 boxes × 3 pH...

Pareto Optimization of Masked Superstrings Improves Compression of Pan-Genome k -Mer Sets

The growing interest in k -mer-based methods across bioinformatics calls for compact k -mer set representations that can be optimized for specific downstream applications. Recently, masked superstrings (MS) have provided such flexibility by moving beyond de Bruijn graph paths to general k -mer superstrings equipped wit...

Graph neural network-based visual design element association analysis and style transfer model

The growing demand for personalized design has driven the development of advanced models for visual design element association and style transfer. However, existing methods struggle to capture complex relationships among design elements and maintain aesthetic coherence during style transfer. The research proposes a nov...

Adaptive heterogeneous graph neural networks for differentiating pulmonary arterial hypertension from left heart disease

Abstract Differentiating pulmonary arterial hypertension (PAH) from pulmonary hypertension associated with left heart disease (PH-LHD) is clinically important because management differs substantially, but definitive classification requires invasive haemodynamic assessment together with clinical evaluation. We retrospec...

Research on dynamic functional connectivity and graph theoretic features of different motor subtypes in Parkinson’s disease based on ICA

Objective This study combined independent component analysis (ICA)-derived dynamic functional connectivity (dFNC) with dynamic graph theory analysis methods to analyze the changes in dFNC and dynamic graph theoretic indicators in different motor subtypes of Parkinson’s disease (PD)—tremor-dominant (TD) subtype and post...