AIscEA

AIscEA integrates single-cell gene expression and chromatin accessibility measurements to enable joint analysis of regulatory relationships between chromatin state and transcription at single-cell resolution.

Key Features:

• Biological consistency-based integration: Quantifies biological consistency between cell clusters across modalities using a ranked similarity score.
• Permutation test for cluster alignment: Applies a novel permutation test to identify significant alignments between heterogeneous clusters across measurements.
• Graph alignment for cell integration: Uses graph alignment techniques to integrate cells across aligned clusters from different modalities.
• Robustness to hyper-parameter selection: Demonstrates stable performance across hyper-parameter choices in unsupervised integration settings.
• Benchmark validation on SNARE-seq and scMultiome-seq: Achieves superior integration accuracy on benchmark datasets including SNARE-seq and scMultiome-seq.

Scientific Applications:

• Enhanced understanding of gene regulation: Facilitates analysis of how chromatin accessibility influences gene expression at single-cell resolution.
• Identification of heterogeneous cell populations: Enables discovery and characterization of diverse cell populations by handling cluster heterogeneity across modalities.
• Robust data integration in single-cell omics studies: Supports integrated analyses of matched gene expression and chromatin accessibility datasets such as SNARE-seq and scMultiome-seq.

Methodology:

AIscEA computes a ranked similarity score based on biological consistency to quantify cluster similarity across gene expression and chromatin accessibility, applies a permutation test to detect significant cluster alignments, and performs graph alignment to integrate cells across measurements.