Celline

Celline executes one-step retrieval and integrative analysis of public single-cell RNA sequencing data through automated workflow execution with single-line commands per step.

Key Features:

• Automated Data Retrieval: Celline automatically gathers raw single-cell RNA-seq data from multiple public repositories.
• Metadata Extraction: Celline extracts metadata using large language models.
• Doublet Removal: Celline wraps Scrublet for doublet removal.
• Quality Control and Cell-Type Annotation: Celline wraps Seurat and Scanpy for quality control and cell-type annotation.
• Batch Correction: Celline wraps Harmony and scVI for batch correction.
• Trajectory Inference: Celline wraps Slingshot for trajectory inference.
• Modular Pipeline Design: The modular design allows pipeline extension and community-driven advances.

Scientific Applications:

• Developmental Biology: Celline was applied to 2 mouse brain cortex datasets from embryonic days 14.5 and 18, annotating 11 major cell types and completing trajectory analysis.
• Disease and Therapy Research: Celline enables integrative analysis of public scRNA-seq datasets that provide insight into development, disease, and therapy.
• Batch Effect Integration: Celline improves integration quality across datasets, demonstrated by a scIB score increase of +0.22.

Methodology:

Celline uses single-line commands per step, wrapping established tools including Scrublet, Seurat, Scanpy, Harmony, scVI, and Slingshot. Technical validation was performed on mouse brain cortex datasets from embryonic days 14.5 and 18.