CoxFormer Documentation

Overview of the CoxFormer framework. CoxFormer first learns transcriptome-wide gene embeddings from co-expression signals. These embeddings are then integrated into a multimodal Transformer-based generative framework for spatial inference, enabling gene-, cell-, and tissue-level downstream tasks.

Tutorials

• Co-expression Completion Tutorial
• Celltype Annotation Tutorial
• Gene Expression Prediction Tutorial
• Gene Activity Score Prediction Tutorial
• Super Resolution Enhancement Tutorial
• Pathological Region Detection Tutorial