CITL

CITL infers time-lagged causal relationships among genes from single-cell RNA sequencing (scRNA-seq) data by leveraging RNA velocity to assess conditional independence between changing and current expression levels.

Key Features:

• Time-lagged causal inference: Infers time-lagged causal relationships among genes using temporal information extracted from scRNA-seq data.
• RNA velocity estimation: Estimates RNA velocity to quantify the rate of change in gene expression and capture dynamic behavior at the single-cell level.
• Conditional independence testing: Assesses conditional independence between changing (velocity) and current expression levels to identify causal links.
• Temporal dependency capture: Detects temporal dependencies and causal relationships that are not apparent in static cross-sectional expression snapshots.
• Benchmarking with simulation: Validated for accuracy and stability using simulation data and compared against other leading approaches, showing superior performance.
• Empirical discovery: Applied to real scRNA-seq datasets and identified 878 pairs of time-lagged causal relationships among genes.

Scientific Applications:

• Inferring gene regulatory networks: Reconstruction of time-lagged regulatory interactions among genes at single-cell resolution.
• Developmental biology: Analysis of temporal gene expression dynamics during development to identify causal regulators.
• Cancer research: Identification of temporal regulatory interactions relevant to cancer progression and heterogeneity.
• Systems biology: Modeling dynamic regulatory networks and temporal dependencies in cellular processes.

Methodology:

Estimate RNA velocity from scRNA-seq to quantify expression change rates and assess conditional independence between changing and current expression levels to infer time-lagged causality; validate using simulation data and apply to real scRNA-seq datasets.