Exploring the Unknown: How Can We Improve Single-cell RNAseq Cell Type Annotations in Non-model Organisms?

Area of Science:

• Genomics
• Bioinformatics
• Evolutionary Biology

Background:

• Single-cell RNA sequencing (scRNAseq) is vital for characterizing cell types across species.
• Standard scRNAseq annotation pipelines, designed for model organisms, struggle with non-model species due to divergent cell types.
• This limitation hinders the discovery of novel species-specific cell types and can lead to mis-annotations.

Purpose of the Study:

• To present novel approaches for annotating single-cell RNA sequencing (scRNAseq) data in non-model organisms.
• To facilitate the discovery of new cell types and improve annotation accuracy in understudied species.
• To enhance the understanding of cellular diversity in the animal kingdom.

Main Methods:

• Utilizing evolutionary context and cell lineage analysis for marker-free cell type annotation.
• Highlighting open-source, reference-free machine learning programs for cluster annotation.
• Proposing the use of unannotated genes as potential cell markers in organisms with incomplete genome annotations.

Main Results:

• Evolutionary context aids in discovering novel cell types and comparing across species.
• Machine learning and reference-free methods offer robust annotation strategies.
• Unannotated genes represent a valuable, often overlooked, resource for cell marker identification.

Conclusions:

• Improved single-cell annotation in non-model organisms is crucial for novel cell type discovery.
• Unified annotation approaches increase confidence in label transfer and flexibility in identifying new cell types.
• This work enhances the cellular-level understanding of non-model organisms.