Unraveling Neuronal Identities Using SIMS: A Deep Learning Label Transfer Tool for Single-Cell RNA Sequencing...

Area of Science:

• Computational Biology
• Genomics
• Machine Learning

Background:

• Single-cell RNA sequencing generates large datasets crucial for biological insight and cell atlases.
• Accurate cell annotation is challenging, especially in complex tissues, limiting the utility of current classification algorithms.

Approach:

• We present SIMS (Scalable, Interpretable Machine Learning for Single-Cell), an end-to-end, data-efficient machine learning pipeline for discrete classification of single-cell data.
• SIMS requires minimal coding and was benchmarked against existing label transfer tools, showing comparable or superior performance.

Key Points:

• SIMS accurately classifies cells in complex tissues like the adult cerebral cortex and hippocampus, maintaining accuracy in trans-sample transfers.
• The tool effectively predicts neuronal subtypes in the developing brain, even during fate refinement, and identifies genetic changes.
• SIMS analyzes cortical organoid datasets, detecting cell-line differences and misannotated lineages, with label transfer from primary tissue improving accuracy.

Conclusions:

• SIMS is a versatile and robust tool for high-accuracy cell-type classification across diverse single-cell datasets.
• The pipeline offers a scalable and interpretable solution for automated cell labeling and biological discovery.