Querying structural and functional niches on spatial transcriptomics data
Researchers have developed QueST, a novel computational method designed to identify similar cellular niches across different spatial transcriptomics samples. This method models niches as subgraphs and utilizes contrastive learning with adversarial training to learn discriminative embeddings and mitigate batch effects. QueST has demonstrated superior performance compared to existing tools in simulations and benchmark datasets, showing accuracy in capturing niche structures and generalizability across sequencing platforms. AI
IMPACT Provides a new tool for analyzing complex biological data, potentially accelerating discoveries in health and disease research.