MuCST

MuCST joins denoising, heterogeneity elimination, and compatible feature learning for integrative analysis of multi-modal spatially resolved transcriptomics (SRT) data using contrastive learning.

Key Features:

• Multi-modal data integration: Integrates spatial location, histology images, and transcriptional profiles from SRT data.
• Denoising: Restores heterogeneous morphology images and spatial transcriptomics data.
• Heterogeneity elimination: Removes heterogeneity across diverse datasets and platforms.
• Compatible feature learning: Learns features that are compatible across different data modalities.
• Spatial domain identification: Accurately identifies spatial domains in undissociated tissues.
• Cross-platform applicability: Applicable to diverse SRT datasets and platforms.

Scientific Applications:

• Spatial domain identification: Identifies spatial domains in multi-modal SRT data from undissociated tissues.
• Integrative analysis of SRT data: Provides an alternative approach for integrative analysis of multi-modal spatially resolved transcriptomics data combining imaging and transcriptomic measurements.

Methodology:

MuCST employs contrastive learning for multi-modal data integration.