Results 131 to 140 of about 81,862 (314)

Spectral Decomposition of Chemical Semantics for Activity Cliffs‐Aware Molecular Property Prediction

open access: yesAdvanced Science, EarlyView.
PrismNet mimics chemical intuition by functioning as a computational prism, refracting molecular graphs into complementary semantic views and spectral frequencies. This dual‐decomposition strategy effectively captures both global topologies and subtle “activity cliff” perturbations.
Chaoyang Xie   +9 more
wiley   +1 more source

CACLENS: A Multitask Deep Learning System for Enzyme Discovery

open access: yesAdvanced Science, EarlyView.
CACLENS, a multimodal and multi‐task deep learning framework integrating cross‐attention, contrastive learning, and customized gate control, enables reaction type classification, EC number prediction, and reaction feasibility assessment. CACLENS accelerates functional enzyme discovery and identifies efficient Zearalenone (ZEN)‐degrading enzymes.
Xilong Yi   +5 more
wiley   +1 more source

Inferring Gene Regulatory Networks From Single‐Cell RNA Sequencing Data by Dual‐Role Graph Contrastive Learning

open access: yesAdvanced Science, EarlyView.
RegGAIN is a novel and powerful deep learning framework for inferring gene regulatory networks (GRNs) from single‐cell RNA sequencing data. By integrating self‐supervised contrastive learning with dual‐role gene representations, it consistently outperforms existing methods in both accuracy and robustness.
Qiyuan Guan   +9 more
wiley   +1 more source

Biomechanics‐Driven 3D Architecture Inference from Histology Using CellSqueeze3D

open access: yesAdvanced Science, EarlyView.
CellSqueeze3D reconstructs 3D cellular architecture from standard 2D histology images using biomechanical constraints and optimization. Validated on clinical datasets, it enables accurate tissue phenotyping, predicts gene mutations, and reveals significant correlations between nuclear‐cytoplasmic ratio entropy and tumor progression.
Yan Kong, Hui Lu
wiley   +1 more source

Multimodal AI‐Driven Identification of Dehydrocostus Lactone as a Potent Renal Fibrosis Attenuator Targeting IQGAP1

open access: yesAdvanced Science, EarlyView.
Renal fibrosis, a hallmark of CKD, lacks effective treatments. Herein, we developed a multimodal AI model (TCM‐SPred) to identify anti‐fibrotic agents and found that dehydrocostus lactone (DCL) targets IQGAP1 to inhibit Wnt signaling, blocking the interaction between IQGAP1 and CCT3, demonstrating potent anti‐fibrotic activity in vitro and in vivo ...
Weijiang Lin   +12 more
wiley   +1 more source

Data cleaning and enrichment through data integration: networking the Italian academia

open access: yesScientific Data
We describe a bibliometric network characterizing co-authorship collaborations in the entire Italian academic community. The network, consisting of 38,220 nodes and 507,050 edges, is built upon two distinct data sources: faculty information provided by ...
Irene Finocchi   +3 more
doaj   +1 more source

MicrobeDiscover: A Knowledge Graph–Enabled AI Framework for Identifying Microbes for Inorganic Nanomaterial Biosynthesis

open access: yesAdvanced Science, EarlyView.
Microbial synthesis of nanomaterials (NMs) is eco‐friendly, but the screening of microorganisms is limited by inefficient traditional methods (currently only involving∽400 microorganisms/90 NMs). We propose AI framework MicrobeDiscover, integrating a knowledge graph of microbe‐NM interactions.
Ludi Wang   +12 more
wiley   +1 more source

Dispersive Full‐Channel Jones Matrix Modulation in Elliptical Polarization Bases via a Single‐Layered Metasurface

open access: yesAdvanced Science, EarlyView.
This work presents a dispersive full‐channel Jones matrix modulation strategy using single‐layer metasurface. By synergizing wavelength dispersion engineering with elliptical polarization bases, independent control of four Jones matrix channels is achieved across multiple wavelengths.
Hairong He   +10 more
wiley   +1 more source

Home - About - Disclaimer - Privacy