Abstract
Astrocytes are considered a diverse cell population, carrying out many functions essential for supporting neuronal activity. The surge of sc/snRNA-sequencing data greatly expands our understanding of heterogeneous astrocyte gene expression, but also leads to confusion about the multitude of described astrocyte subtypes and substates in the mammalian brain. Here we discuss and review the definition of distinct subtypes and the evidence for this amongst astrocytes. Determining whether an astrocyte subtype represents a stable identity or a dynamic substate requires generalization of findings across datasets, incorporation of validation, and ideally, functional analyses. How to best achieve this is the focus of this review, including considerations about the different transcriptomic approaches. We further discuss the alignment of astrocyte subtype transcriptomes with other hallmarks, such as their position. These considerations are embedded in an overview of the current astrocyte heterogeneity knowledge as a basis for subtype definitions using different analysis techniques. Following technical and biological considerations of transcriptome analyses, we advocate for multimodal alignment to identify stable astrocyte subtypes.