stCluster

stCluster integrates graph contrastive learning with multi-task learning to decipher spatial domains in spatially resolved transcriptomics data, improving spatial domain identification and reducing noise in spatial gene expression patterns.

Key Features:

• Graph Contrastive Learning: stCluster leverages graph contrastive learning technology to obtain discriminative representations capable of recognizing spatially coherent patterns.
• Multi-Task Learning: Through jointly optimizing multiple tasks, stCluster fine-tunes representations to capture complex relationships between gene expression and spatial organization.
• Denoising: stCluster effectively denoises spatial gene expression patterns.
• Spatial Trajectory Inference Enhancement: stCluster enhances spatial trajectory inference.

Scientific Applications:

• Spatial Domain Identification: stCluster accurately identifies complex spatial domains across various datasets and platforms, spanning tissue, organ, and embryo levels.
• Spatial Gene Expression Denoising: stCluster reduces noise in spatial gene expression patterns.
• Spatial Trajectory Inference: stCluster enhances inference of spatial trajectories in tissue contexts.

Methodology:

stCluster first leverages graph contrastive learning to obtain discriminative representations for recognizing spatially coherent patterns, then jointly optimizes multiple tasks to fine-tune representations that capture relationships between gene expression and spatial organization.