AS‐pHopt: An Optimal pH Prediction Model Enhanced by Active Site of Enzymes
Advanced Intelligent Discovery, EarlyView.To address the low accuracy of enzyme optimal pH (pHopt) prediction, this study develops active site‐based pHopt (AS‐pHopt), a prediction model enhanced by active site information and pseudo‐label prediction. Integrating key structural and physicochemical features affecting enzyme pHopt, AS‐pHopt uses Evolutionary Scale Modeling (ESM)‐2 with active ...
Wenxiang Song +6 more
wiley +1 more source
DWGCN: distance-weighted graph convolutional network for robust spatial domain identification in spatial transcriptomics. [PDF]
Front GenetPeng C, Li G, Wu J, Fan Q, Guo X.
europepmc +1 more source
Advanced Intelligent Discovery, EarlyView.We report a novel interpretation method for deep learning models based on feature extraction and clustering. Applying this method to an atomistic line graph neural network (ALIGNN) model trained on optical absorption spectra of 2,681 inorganic compounds obtained from first‐principles calculations, we successfully identify key factors underlying ...
Akira Takahashi +3 more
wiley +1 more source
Advanced Intelligent Discovery, EarlyView.AI‐BioMech is a deep learning framework that predicts the mechanical behavior of biological cellular materials directly from 2D images. By replacing traditional finite element analysis with semantic segmentation, it identifies stress and strain distributions with 99% accuracy, offering a high‐speed, scalable alternative for analyzing complex, aperiodic
Haleema Sadia +2 more
wiley +1 more source
STransfer: a transfer learning-enhanced graph convolutional network for clustering spatial transcriptomics data. [PDF]
BioinformaticsWang C, Yu X.
europepmc +1 more source
Multimodal Learning with Rashomon Analysis for Battery Discharge Capacity Prediction
Advanced Intelligent Discovery, EarlyView.Multimodal fusion integrates composition, crystal‐structure, and radial‐distribution descriptors to predict battery discharge capacity. Rashomon analysis across near‐optimal models reveals that explanatory variation is structured rather than arbitrary, separating stable mechanistic signals from model‐contingent attributions and providing a more ...
Jue Gong +4 more
wiley +1 more source
Innovative application of a traffic-prediction spatio-temporal graph convolutional network for dengue disease forecasting. [PDF]
Sci RepSiabi N +4 more
europepmc +1 more source
Materials Representation Learning Based on a Material–Motif Network and Heterogeneous Graphs
Advanced Intelligent Discovery, EarlyView.Structure motifs in materials are used to construct a bipartite material–motif network that links each material to its constituent motifs and establishes connectivity among materials sharing common motifs. Network analysis reveals material clusters associated with different functional applications and supports motif‐guided screening of materials.
Anoj Aryal +3 more
wiley +1 more source
OptiSyn: an interpretable, multi-omics-driven graph convolutional network framework for synergy-oriented drug combination design in disease treatment. [PDF]
Chin MedShi Y +9 more
europepmc +1 more source
Advanced Intelligent Discovery, EarlyView.Machine learning serves as a central engine for the intelligent characterization of two‐dimensional materials by integrating multimodal techniques, including optical microscopy, spectroscopy, electron microscopy, and scanning probe microscopy (SPM). This unified framework enables automated, high‐throughput, and quantitative extraction of structural ...
Zhi‐Long Cao, Jia‐Xu Yan
wiley +1 more source