Results 131 to 140 of about 211,846 (266)
Advanced Science, EarlyView.HiST, a multiscale deep learning framework, reconstructs spatially resolved gene expression profiles directly from histological images. It accurately identifies tumor regions, captures intratumoral heterogeneity, and predicts patient prognosis and immunotherapy response.
Wei Li +8 more
wiley +1 more source
Cancer‐Like Fragmentomic Characteristics of Somatic Variants in Cell‐Free DNA
Advanced Science, EarlyView.We report that in non‐cancerous subjects, cell‐free (cfDNA) molecules harboring somatic variants exhibit cancer‐like fragmentomic characteristics associated with clonal hematopoiesis. Importantly, these somatic variant‐associated fragmentomic aberrations are more pronounced in cancer patients. Leveraging such somatic variant‐associated signals in cfDNA,
Zhenyu Zhang +12 more
wiley +1 more source
OBUSight: Clinically Aligned Generative AI for Ophthalmic Ultrasound Interpretation and Diagnosis
Advanced Science, EarlyView.OBUSight, a clinically aligned generative AI model that jointly generates reports and predicts diseases through multimodal semantic alignment, was trained and validated on a large multicenter dataset. OBUSight outperformed eight state‐of‐the‐art models, provided clinically reliable reports, enhanced diagnostic efficiency, and achieved performance ...
Xiaocong Liu +17 more
wiley +1 more source
Advanced Science, EarlyView.ImmuDef, a novel algorithm to quantitatively evaluate the anti‐infection immune defense function of an individual based on RNA‐seq data via a variational autoencoder (VAE) model. It is validated on 3200+ samples across four immune states with high accuracy. It can serve as a metric for disease severity and prognosis across pathogenic cohorts.
Zhen‐Lin Tan +7 more
wiley +1 more source
Advanced Science, EarlyView.This study develops a deep learning‐based pathomics model to predict survival outcomes in pancreatic cancer patients. The CrossFormer architecture analyzes routine H&E‐stained tissue slides, identifying key prognostic features including stromal patterns, cellular characteristics, and immune infiltration.
Qiangda Chen +22 more
wiley +1 more source
S3RL: Enhancing Spatial Single‐Cell Transcriptomics With Separable Representation Learning
Advanced Science, EarlyView.Separable Spatial Representation Learning (S3RL) is introduced to enhance the reconstruction of spatial transcriptomic landscapes by disentangling spatial structure and gene expression semantics. By integrating multimodal inputs with graph‐based representation learning and hyperspherical prototype modeling, S3RL enables high‐fidelity spatial domain ...
Laiyi Fu +6 more
wiley +1 more source
Accurate Identification of Protein Binding Sites for All Drug Modalities Using ALLSites
Advanced Science, EarlyView.ALLSites is a unified sequence‐based framework for identifying proteome‐wide binding sites across all drug modalities. It integrates a gated convolutional network with a transformer architecture to capture residue interactions directly from the sequence.
Minjie Mou +14 more
wiley +1 more source
Overtuning in Hyperparameter Optimization
Accepted at the Fourth Conference on Automated Machine Learning (Methods Track).
Schneider, Lennart +2 more
openaire +2 more sources
Advanced Science, EarlyView.ABSTRACT Precise transgene‐free gene upregulation remains a challenge in crop biotechnology, as conventional enhancers often exceed CRISPR‐mediated knock‐in size constraints and face regulatory hurdles. Here we establish a foundational cross‐species resource of compact transcriptional enhancers developed via STEM‐seq, a high‐throughput screening ...
Qi Yao +14 more
wiley +1 more source