Detection of Spatially Aberrant Cells in Spatial Transcriptomics Data by Conformal Prediction
Zhang, Z.; Zheng, X.; Yuan, Q.; Luo, R.
Show abstract
The hexagonal organization of epithelial cells represents a fundamental feature of normal tissue architecture, reflecting the precise spatial coordination that underlies healthy biological structure. Disruptions to this organization--manifesting as spatially aberrant spots with abnormal gene expression and misplaced positioning--are closely associated with disease initiation and progression. Here, we introduce SPADE, a computational framework that integrates single-cell RNA sequencing and spatial transcriptomics data to quantitatively characterize and detect spatial aberrancy. SPADE leverages a variational autoencoder coupled with Gaussian mixture modeling for cell-type embedding and spatial deconvolution, and incorporates conformal prediction to enable uncertainty-calibrated identification of aberrant spots. Through extensive validation, SPADE demonstrates superior performance in identifying biologically meaningful aberrant spots.
Matching journals
The top 4 journals account for 50% of the predicted probability mass.
Similar papers in this journal
Similar papers in this journal
- Learning multi-cellular representations of single-cell transcriptomics data enables characterization of patient-level disease states 96%
- Identifying maximally informative signal-aware representations of single-cell data using the Information Bottleneck 95%
- Single-cell morphodynamical trajectories enable prediction of gene expression accompanying cell state change 95%
Similar papers in this journal
- Information-Distilled Generative Label-Free Morphological Profiling Encodes Cellular Heterogeneity 93%
- DeDoc2 identifies and characterizes the hierarchy and dynamics of chromatin TAD-like domains in the single cells 92%
- Unraveling the molecular mechanisms underlying spontaneous multipolar mitosis through CIN-seq 92%
Similar papers in this journal
- Geometry aware graph attention networks to explain single-cell chromatin state and gene expression 96%
- Neighborhood nonnegative matrix factorization identifies patterns and spatially-variable genes in large-scale spatial transcriptomics data 94%
- Explainable multi-view framework for dissecting inter-cellular signaling from highly multiplexed spatial data 94%
"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.