Characterizing the replicability of cell types defined by single cell RNA-sequencing data using MetaNeighbor

Area of Science:

• Genomics
• Computational Biology
• Neuroscience

Background:

• Single-cell RNA sequencing (scRNA-seq) is powerful for cell type discovery.
• Technical biases and inconsistent naming conventions hinder scRNA-seq data replicability and meta-analysis.

Purpose of the Study:

• To develop and validate a framework (MetaNeighbor) for quantifying cell type replication across diverse scRNA-seq datasets.
• To establish best practices for assessing cell type reproducibility.
• To identify robust interneuron subtypes and candidate marker genes.

Main Methods:

• Developed MetaNeighbor, a computational framework to measure cell type similarity across datasets.
• Applied MetaNeighbor to eight diverse scRNA-seq datasets, initially focusing on neuronal identity.
• Validated the framework on novel interneuron subtypes.

Main Results:

• MetaNeighbor accurately quantifies cell type replication across datasets.
• 24 out of 45 novel interneuron subtypes showed evidence of replication.
• Large sets of variably expressed genes effectively identify replicable cell types.

Conclusions:

• MetaNeighbor enhances the replicability and meta-analysis of scRNA-seq data.
• The framework facilitates the identification of robust cell types and candidate marker genes.
• This approach offers a generalizable strategy for large-scale scRNA-seq data evaluation.