scWECTA: A weighted ensemble classification framework for cell type assignment based on single cell transcriptome.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Single-cell RNA sequencing (scRNA-seq) reveals cellular heterogeneity but requires accurate cell type annotation.
• Current automated annotation methods often rely on single feature sets and classifiers, limiting their performance.
• Identifying cell types, especially rare ones, remains a challenge in scRNA-seq data analysis.

Purpose of the Study:

• To develop an improved computational framework for accurate cell type annotation in scRNA-seq data.
• To address limitations of existing methods by integrating multiple feature sets and classifiers.
• To enhance the identification of both common and rare cell populations.

Main Methods:

• Proposed scWECTA (Weighted Ensemble classification framework for Cell Type Annotation).
• Integrated five informative gene sets and five distinct classifiers using a soft weighted ensemble approach.
• Employed constrained non-negative least squares for inferring ensemble weights.

Main Results:

• scWECTA demonstrated high accuracy in annotating scRNA-seq data across different platforms and tissues.
• The method showed superior performance on imbalanced datasets, effectively identifying rare cell types.
• scWECTA balanced the prediction accuracy of common cell types with a low unassigned rate for non-common cell types.

Conclusions:

• scWECTA offers a robust and accurate solution for cell type annotation in scRNA-seq data.
• The weighted ensemble approach effectively improves annotation accuracy, particularly for challenging datasets.
• The freely available source code facilitates broader application in single-cell research.