Attention-based framework for cell-cell communication analysis

This technology is a software that analyzes cell-cell interactions to uncover gene programs and signaling nodes in disease.

Unmet Need: Capturing dynamic, context-dependent cell-cell communication

Spatial transcriptomics measures gene expression at defined spatial coordinates within tissue sections, enabling analysis of cell–cell communication within intact tissues. However, current computational methods mainly provide static snapshots and do not capture how interactions change across different microenvironments or time points. As such, developing approaches that can capture the dynamic and context-dependent nature of cellular communication is crucial.

The Technology: An attention-based software for determining cell-cell interactions

This technology, termed AMICI (Attention Mechanism Interpretation of Cell-cell Interactions), is an attention-based framework for dynamic analysis of cell-cell communications. This software provides sender-receiver subpopulation resolution, adaptive modeling of interaction length scales, and interpretable attention mechanisms. This can help uncover gene programs and communication signaling nodes associated with disease.

Applications:

• Mapping tissue architecture
• Tumor microenvironment profiling
• Developmental biology and regeneration studies
• Integrative multi-omics studies

Advantages:

• Dynamic and temporal cell-cell interaction information
• Attention-based framework
• Captures tissue heterogeneity
• Improves biological interpretability

Lead Inventor:

Elham Azizi, Ph.D.

Patent Information:

Patent Pending

Related Publications:

• Park C, Mani S, Beltran-Velez N, Maurer K, Huang T, Li S, Gohil S, Livak KJ, Knowles DA, Wu CJ, Azizi E. “A Bayesian framework for inferring dynamic intercellular interactions from time-series single-cell data.” Genome Res. 2024 Oct 11;34(9):1384-1396.

Tech Ventures Reference:

• IR CU26099

• Licensing Contact: Joan Martinez