Protocol for deep-learning-driven cell type label transfer in single-cell RNA sequencing data.

Area of Science:

• Computational Biology
• Genomics
• Machine Learning

Background:

• Single-cell RNA sequencing (scRNA-seq) generates high-dimensional data crucial for understanding cellular heterogeneity.
• Accurate cell type annotation is a fundamental challenge in scRNA-seq data analysis.
• Transferring cell type labels can accelerate analysis and enable cross-dataset comparisons.

Purpose of the Study:

• To present a standardized protocol for utilizing SIMS for cell type label transfer in scRNA-seq data.
• To provide a user-friendly framework for data preparation, model training, and prediction interpretation.
• To facilitate the accessibility of advanced machine learning tools for scRNA-seq analysis.

Main Methods:

• Development of a protocol for SIMS, a machine learning framework for single-cell data.
• Stepwise instructions for data preparation, including handling and preprocessing.
• Model training using labeled datasets or inference with pre-trained models for cell type annotation.

Main Results:

• A comprehensive protocol detailing the application of SIMS for cell type label transfer.
• Methods for visualizing, downloading, and interpreting the predicted cell type labels.
• Demonstration of SIMS accessibility via API, GitHub Codespaces, and a web application.

Conclusions:

• The SIMS protocol offers a robust and accessible method for cell type annotation in scRNA-seq data.
• This approach enhances the interpretability and scalability of machine learning in single-cell genomics.
• The provided resources empower researchers to efficiently transfer cell type labels and advance biological insights.