Technical limitations often let dominant signals overshadow rare cell types and fine‐grained heterogeneity in spatial transcriptomics. SemanticST, a graph neural network using multi‐semantic graph fusion and a novel min‐cut loss, recovers these subtle patterns.
Roxana Zahedi +7 more
wiley +1 more source
Riemannian metric learning for alignment of spatial multiomics. [PDF]
Halmos P, Xia Y, Raphael BJ.
europepmc +1 more source
This study presents BraMARS, an explainable deep learning model that estimates future brain metastasis risk in surgically resected limited‐stage small‐cell lung cancer using routine H&E‐stained whole‐slide images. By linking model‐attributed spatial histopathology with clinical outcomes and proteomic programs, BraMARS provides a biologically ...
Zijian Yang +10 more
wiley +1 more source
Full dynamic similarity experiments and predictive modeling of wind turbine aerodynamics across control strategies. [PDF]
Kurelek JW +6 more
europepmc +1 more source
Dipolar Coupling Errors in Nuclear Spin Qubits
Multi‐nuclear spin registers of phosphorus‐atom qubits in silicon have demonstrated 99.99% single and >$>$99.9% two qubit gate fidelities. Despite nanometer separation, read‐out fidelities between individual nuclear spins within the register are not identical.
I. D. Thorvaldson +15 more
wiley +1 more source
Application of novel similarity measures in electric vehicle charging station site selection based on q-rung orthopair hesitant fuzzy rough sets under quantitative information. [PDF]
Attaullah +4 more
europepmc +1 more source
Computationally Evidence‐Grounded Sequence‐First Design of Peptide Binders
BOND‐PEP enables controllable, sequence‐first peptide binder design by grounding generation in binding evidence retrieved for each target. It uses topology‐conditioned message passing to integrate relevant peptide examples with the target protein sequence, forming a residue‐level representation that guides the generation of diverse, target‐specific ...
Wenze Ding
wiley +1 more source
A unified framework of information measures for complex linear Diophantine fuzzy sets with application to chronic kidney disease. [PDF]
Mustafa AW, Bashir Z, Ali J, Syam MI.
europepmc +1 more source
Learning Work Function via Implicit Reasoning on Electrostatic Potential Landscapes
StructPot‐CLR establishes a cross‐modal contrastive learning framework that aligns the crystal structures of 2D materials with plane‐averaged electrostatic potential landscapes for physically informed work‐function prediction. The model achieves an MAE of 0.265 eV and an R2 of 0.902 on the held‐out test set while accurately preserving key morphological
Haoyu Wan, Yue Wu, Tianhao Su, Deng Pan
wiley +1 more source
An embedding-based framework enables statistical testing of gene-set function hypotheses inferred by large language models. [PDF]
Tan Y +8 more
europepmc +1 more source

