Accelerating scRNA-seq Analysis: Automated cell type annotation using representation learning and vector search

Area of Science:

• Genomics
• Computational Biology
• Bioinformatics

Background:

• Cell type annotation is essential for interpreting single-cell RNA sequencing (scRNA-seq) data.
• Accurate cell categorization is critical for extracting meaningful biological insights from scRNA-seq experiments.

Purpose of the Study:

• To develop and present an automated service for cell type annotation of 10x Genomics scRNA-seq data.
• To enable researchers to quickly and accurately assign cell types within their samples.

Main Methods:

• The service employs a reverse search strategy, comparing individual cell gene expression profiles.
• It utilizes the Chan Zuckerberg CELL by GENE (CZ CELLxGENE) Census, a repository of annotated scRNA-seq datasets.
• Annotations are generated by summarizing cell type labels from similar cells in the reference dataset.

Main Results:

• The service provides automated cell type annotations for scRNA-seq samples.
• It offers both fine-grained and coarse-level annotations.
• The annotation process does not rely on predefined marker genes or tissue-specific references.

Conclusions:

• The automated service facilitates rapid and accurate cell type annotation in scRNA-seq experiments.
• This tool empowers researchers to accelerate biological discovery by efficiently categorizing cells.
• The generated annotations can be further refined by users for specific research applications.