Rectangle

Rectangle combines multiscale deconvolution with explicit modeling of unknown cellular content to enable single-cell-informed deconvolution of bulk RNA-seq and spatial transcriptomic data.

Key Features:

• Hierarchical signature building: Enables fine-grained cell-type deconvolution through multiscale resolution levels.
• Unknown cellular content modeling: Estimates and corrects for cellular content not represented in the single-cell reference.
• Scalability: Efficiently handles large-scale single-cell data during signature matrix computation.
• Cross-method benchmark performance: Achieves high accuracy, high resolution, low spillover, strong scalability and efficiency, and robustness to unknown cellular content.

Scientific Applications:

• Cell-type and cell-state profiling at scale: Bridges single-cell transcriptomics resolution with bulk RNA-seq scale and cost-efficiency for population-scale studies.
• Cellular biomarker discovery: Supports identification of cellular biomarkers in large cohorts.
• Tracking cellular dynamics: Enables monitoring of cellular changes in settings impractical for comprehensive single-cell sequencing.

Methodology:

Rectangle performs hierarchical signature building for fine-grained cell-type deconvolution, explicitly models unknown cellular content, and computes signature matrices from large-scale single-cell data.