Cisformer

Cisformer leverages cross-attention to link cis-regulatory elements to target genes, facilitating cross-modality generation between gene expression and chromatin accessibility at single-cell resolution and the identification of functional transcription factors.

Key Features:

• Cross-attention mechanism: The model uses cross-attention to precisely link cis-regulatory elements to target genes, providing inherent biological interpretability.
• Cross-modality generation: Cisformer generates one modality from another between gene expression and chromatin accessibility at single-cell resolution.
• Superior accuracy and generalization: Systematic benchmarking demonstrates Cisformer outperforms existing methods in accuracy and generalization.

Scientific Applications:

• Functional transcription factor identification: Cisformer identifies functional transcription factors associated with tumorigenesis and aging.
• Single-cell multiomic data analysis: Cisformer enables analysis of single-cell multiomic data through cross-modality translation.

Methodology:

Cisformer is a cross-attention-based generative model tailored for cross-modality generation between gene expression and chromatin accessibility at single-cell resolution.