Reproduce “stMCDI: Masked Conditional Diffusion Model with Graph Neural Network for Spatial Transcriptomics Data Imputation” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Xiaoyu Li, Wenwen Min, Shunfang Wang, Changmiao Wang, Taosheng Xu
stMCDI introduces a masked conditional diffusion model with a graph neural network to impute missing gene expression values in spatial transcriptomics data, leveraging spatial context and preserving data distribution. The method consistently outperformed fourteen baseline techniques on six real-world datasets across multiple quantitative metrics.
Key insights
- Explicitly embedding spatial context into gene expression data via a GNN encoder significantly enhances the accuracy of missing value imputation.
- Conditional diffusion models are effective for generating plausible missing data while maintaining the fidelity of the overall gene expression distribution in complex biological datasets.
- A multi-stage self-supervised masking approach provides a robust framework for training sophisticated imputation models when ground truth labels are unavailable.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.