SCIBER

SCIBER removes batch effects from single-cell RNA-sequencing (scRNA-seq) datasets to enable accurate, interpretable integration of gene-level expression data across batches and studies.

Key Features:

• Batch-effect removal: Removes batch effects from scRNA-seq data to prevent technical variation from confounding biological interpretation.
• Gene-level expression output: Produces corrected expression data in the original gene space rather than only dimension-reduced representations.
• Scalability and simplicity: Implements a simple, scalable algorithm suitable for large numbers of cells.
• Reference-based methodology: Assigns one batch as the reference and keeps its data unchanged during correction, enabling integration with reference datasets such as the Human Cell Atlas.
• Interpretability: Provides corrections that are interpretable at the gene level.
• Performance and accuracy: Demonstrates comparable or superior accuracy to state-of-the-art methods on real datasets.

Scientific Applications:

• Integrative scRNA-seq analysis: Enables joint analysis of multiple scRNA-seq datasets by removing technical batch differences.
• Cross-study and cross-condition comparisons: Supports accurate comparisons across different studies and experimental conditions.
• Cell-type and state characterization: Facilitates comprehensive characterization of cell types and cellular states from integrated gene-level data.
• Domain-specific research: Applicable to developmental biology, immunology, oncology, and regenerative medicine where cellular heterogeneity and dynamics are studied.

Methodology:

Reference-based batch correction that assigns one batch as the reference and keeps its data unchanged, producing corrected gene-level expression in the original gene space using a simple, scalable algorithm.