Monet-RNA

Monet-RNA constructs a PCA-based latent space from scRNA-Seq datasets to infer dataset dimensionality and enable downstream analyses such as batch correction, label transfer, and dataset integration.

Key Features:

• Dimensionality inference: Infers the intrinsic dimensionality of scRNA-Seq datasets to guide dimensionality reduction and analysis.
• PCA-based latent space: Builds a PCA-derived latent space capturing principal axes of variance for downstream analysis.
• MonetModel object: Encapsulates the constructed latent space within a MonetModel object for storage and reuse.
• Batch correction: Facilitates batch correction workflows using the PCA-based latent representation.
• Label transfer: Enables label transfer between datasets via the shared latent space.
• Method development framework: Provides a modular foundation for developing new scRNA-Seq analysis methods.
• Computational efficiency: Implements computationally efficient algorithms suitable for large-scale single-cell data.

Scientific Applications:

• Single-cell tissue analysis: Characterizing cellular heterogeneity and transcriptomic profiles in complex tissues at single-cell resolution.
• Batch correction: Harmonizing datasets from different experimental batches to ensure comparability across studies.
• Label transfer: Transferring annotations between reference and query scRNA-Seq datasets to apply known labels to new data.
• Dataset integration and comparison: Integrating and comparing multiple datasets from the same tissue type using a shared latent space.

Methodology:

Monet-RNA uses PCA-based latent space construction and algorithms to infer dataset dimensionality, and stores the resulting representation in a MonetModel object for downstream analyses such as batch correction and label transfer.