RCSL

RCSL implements a rank constrained similarity learning algorithm to identify cell types from single-cell RNA sequencing data by integrating local and global cell similarity measures.

Key Features:

• Global Similarity Measurement: Uses Spearman's rank correlations of a cell's expression vector with those of other cells to measure global similarity.
• Local Similarity Learning: Adaptively learns neighbour representation of a cell as its local similarity.
• Integrated Similarity Scoring: Computes overall cell similarity as a linear combination of global similarity and local similarity.
• Dual-Sensitivity Clustering: Discerns subtle differences among cells of the same type and larger differences among cells of different types.

Scientific Applications:

• Cell Type Identification: Accurately identifies various cell types using scRNA-seq data from complex tissues.

Methodology:

RCSL uses Spearman's rank correlations for global similarity, adaptively learns neighbour representations for local similarity, and combines these as a linear combination for overall similarity-based clustering.