Spatial Transcriptomics Prediction from Histology Images at Single-cell Resolution using RedeHist
Zhong, Y.; Zhang, J.; Ren, X.
Show abstract
Spatial transcriptomics (ST) offers substantial promise in elucidating the tissue architecture of biological systems. However, its utility is frequently hindered by constraints such as high costs, time-intensive procedures, and incomplete gene readout. Here we introduce RedeHist, a novel deep learning approach integrating scRNA-seq data to predict ST from histology images at single-cell resolution. Application of RedeHist to both sequencing-based and imaging-based ST data demonstrated its outperformance in high-resolution and accurate prediction, whole-transcriptome gene imputation, and fine-grained cell annotation compared with the state-of-the-art algorithms.
Matching journals
The top 5 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Predicting Spatially Resolved Gene Expression via Tissue Morphology using Adaptive Spatial GNNs 95%
- Digitally Predicting Protein Localization and Manipulating Protein Activity in Fluorescence Images Using Four-dimensional Reslicing GAN 95%
- scFeatures: Multi-view representations of single-cell and spatial data for disease outcome prediction 94%
Similar papers in this journal
Similar papers in this journal
- Efficient prediction of a spatial transcriptomics profile better characterizes breast cancer tissue sections without costly experimentation 95%
- Spatial Transcriptomics Inferred from Pathology Whole-Slide Images Links Tumor Heterogeneity to Survival in Breast and Lung Cancer 94%
- Predicting cell types with supervised contrastive learning on cells and their types 94%
Similar papers in this journal
- A Message Passing Framework for Precise Cell State Identification with scClassify2 96%
- RNA2seg: a generalist model for cell segmentation in image-based spatial transcriptomics 96%
- BERMUDA: A novel deep transfer learning method for single-cell RNA sequencing batch correction reveals hidden high-resolution cellular subtypes 95%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.