scME

scME implements a dual-modality factor model to disentangle and integrate shared and complementary molecular features from single-cell multiomics technologies for joint embedding and improved cell clustering and classification.

Key Features:

• Integration of Shared and Complementary Information: Disentangles and integrates both shared signals and modality-specific (complementary) molecular features across modalities.
• Deep Factor Modeling: Uses a deep factor modeling approach to represent latent factors underlying multiple single-cell modalities.
• Joint Representation: Generates a comprehensive joint embedding of multiple modalities to capture nuanced differences among cells.
• Enhanced Clustering and Classification: Produces embeddings that improve single-cell clustering and cell-type classification accuracy.

Scientific Applications:

• Cell heterogeneity analysis: Integrates multiomic features to dissect cellular heterogeneity within complex tissues.
• Developmental biology studies: Resolves subtle cell-state differences relevant to developmental processes.
• Disease mechanism investigation: Enables profiling of cell-type–specific molecular changes in disease contexts.
• Personalized medicine: Supports detailed cellular profiling that can inform individualized biological and clinical interpretations.

Methodology:

scME applies a dual-modality deep factor model to disentangle shared and complementary signals and produce joint embeddings for single-cell multiomics data.