cellHarmony

cellHarmony aligns and classifies single-cell RNA-Seq transcriptomes to enable comparative analysis of cell-type-specific gene expression across conditions for studying molecular pathogenesis.

Key Features:

• Community-clustering and alignment strategy: Employs a community-clustering approach combined with alignment strategies to match single-cell transcriptomes across potentially dozens of cell populations and compute cell-type-specific expression differences.
• Identification of gene programs and pathways: Leverages transcriptional differences to automatically identify distinct and shared gene programs and to pinpoint impacted pathways and transcriptional regulatory networks.
• Integration with AltAnalyze and Python implementation: Implemented as a standalone Python package and as an integrated workflow within AltAnalyze.
• Performance benchmarking: Demonstrated improved or equivalent performance compared to alternative label projection methods.
• Cellular origin identification: Capable of identifying likely cellular origins of malignant states.
• Clinical stratification: Facilitates stratification of patients into clinical disease subtypes based on identified gene programs.
• Disease network resolution: Resolves discrete disease networks impacting specific cell types.
• Therapeutic insights: Reveals how different treatments may affect cellular and molecular pathways to illuminate therapeutic mechanisms.

Scientific Applications:

• Disease mechanism discovery: Enables detailed comparison of single-cell transcriptomes to uncover molecular and cellular origins of complex diseases.
• Patient stratification: Supports stratification of patients into clinical subtypes using cell-type-specific gene programs.
• Therapeutic target identification: Aids identification of potential therapeutic targets and evaluation of treatment effects on cellular pathways.
• Regulatory network analysis: Facilitates elucidation of transcriptional regulatory networks and impacted pathways within specific cell types.
• Cancer origin inference: Assists in inferring the likely cellular origins of malignant states.

Methodology:

Uses community-clustering and alignment strategies to match single-cell RNA-Seq transcriptomes, computes differences in cell-type-specific gene expression, and leverages transcriptional differences to automatically identify distinct/shared gene programs and impacted pathways and transcriptional regulatory networks.