SCIntRuler

SCIntRuler provides a statistical metric to guide the integration of multiple single-cell RNA-seq datasets, assessing the need for integration and informing method selection.

Key Features:

• Integration necessity assessment: SCIntRuler evaluates whether data integration is required or if simple merging of datasets is sufficient for joint analysis.
• Integration method selection: SCIntRuler guides selection of the most appropriate integration method among existing methods for specific scenarios.
• Statistical metric: SCIntRuler introduces a novel statistical metric for quantifying integration needs and comparing method suitability.
• Simulation validation: SCIntRuler is validated through simulations and real data applications across diverse scRNA-seq datasets and varying contexts.

Scientific Applications:

• Multi-study scRNA-seq integration: SCIntRuler facilitates the analysis of heterogeneous single-cell RNA-seq datasets from multiple studies by determining optimal integration strategies.
• Batch effect evaluation: SCIntRuler helps researchers decide whether observed batch discrepancies within or across samples warrant formal integration methods versus simple dataset merging.
• Heterogeneous dataset analysis: SCIntRuler supports analysis of diverse scRNA-seq datasets with inherent diversity and varying batch effects.

Methodology:

SCIntRuler is implemented as an R package. The method uses a novel statistical metric to assess integration necessity and compare integration methods, validated through simulations and real data applications.