Precise cell recovery by cell nucleus united transcript (CellCUT) for enhanced spatial transcriptomics
Wang, X.; Hong, B.; Zeng, B.; Feng, H.; Liu, Z.; Ni, Q.; Wang, W.; Li, M.; Yang, M.; Wang, M.; Sun, L.; ZHong, S.; Wu, Q.
Show abstract
Cell segmentation is the first step in parsing spatial transcriptomic data, often a challenging task. Existing cell segmentation methods do not fully leverage spatial cues between nuclear images and transcripts, tending to produce undesirable cell profiles for densely packed cells. Here, we propose CellCUT to perform cell segmentation and transcript assignment without additional manual annotations. CellCUT provides a flexible computational framework that maintains high segmentation accuracy across diverse tissues and spatial transcriptomics protocols, showing superior capabilities compared to state-of-the-art methods. CellCUT is a robust model to deal with undesirable data such as low contrast intensity, localized absence of transcripts, and blurred images. CellCUT supports a human-in-the-loop workflow to enhance its generalizability to customized datasets. CellCUT identifies subcellular structures, enabling insights at both the single-cell and subcellular levels.
Matching journals
The top 3 journals account for 50% of the predicted probability mass.
Similar papers in this journal
Similar papers in this journal
- stGCL: A versatile cross-modality fusion method based on multi-modal graph contrastive learning for spatial transcriptomics 97%
- High-precision cell-type mapping and annotation of single-cell spatial transcriptomics with STAMapper 97%
- A Systematic Evaluation of Single-cell RNA-sequencing Imputation Methods 96%
Similar papers in this journal
- Information-Distilled Generative Label-Free Morphological Profiling Encodes Cellular Heterogeneity 96%
- Predicting MammaPrint Recurrence Risk from Breast Cancer Pathological Images Using a Weakly Supervised Transformer 96%
- DeDoc2 identifies and characterizes the hierarchy and dynamics of chromatin TAD-like domains in the single cells 95%
Similar papers in this journal
"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.