Back

SpaDiff: Denoising for Sequence-based Spatial Transcriptomics via Diffusion Process

Cai, J.; Chen, Y.; Fang, L.; Zhong, W.; Yuan, G.-C.; Ma, P.

2025-10-08 bioinformatics
10.1101/2025.10.07.681011 bioRxiv
Show abstract

Spatial transcriptomics enables transcriptome-scale analysis with spatial resolution but suffers from spot-swapping, where RNA molecules drift from their true locations, introducing noise and reducing spatial specificity. We introduce SpaDiff, a denoising method that treats spot-swapping as a diffusion process. SpaDiff simulates the displacement of RNA molecules and reverses it to restore their original spatial distribution while preserving molecular counts. Evaluations on simulated and real data demonstrate that SpaDiff enhances spatial specificity of gene expression, improves data integrity, and supports more accurate downstream analyses such as clustering and spatial domain identification.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"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.