What is a cell type and how to define it
Setting up standardized definitions and classification methods for cell types is necessary for reproducible research. This review looks at cell types and states in the brain context. Current methods involve single-cell transcriptomics, spatial transcriptomics, and single-cell epigenomics (ATAC-seq), connectomics where connections between neurons are studied.
Single-cell transcriptomics provide comprehensiveness and high dimensionality to create cell atlases. Single-cell epigenomics with single-nucleus ATAC-seq or methylation sequencing show gene and chromatin regulatory landscapes. Spatial transcriptomics provides information on the distribution of cell and their spatial relationships. The improvement of single-cell proteomics will enhance our understanding of the links between gene expression, cell structure and function, and the relationships between the transcriptome and proteome.
Cell types as a product of evolution
One approach to classify cells is phylogenetic classification which can provide a classification framework based on relatedness in a similar way to the species tree-of-life classification system. This shows the evolutionary distance between cell types and the evolution of new cell type regulatory signatures. These signatures encoded in the genome can be identified in transcriptomic data as shown by the way single-cell transcriptomic clustering reveals cell types consistent with existing knowledge. This data can also be used across species to compare cell types and their evolutionary trajectories and relationships.
Hierarchical organization of transcriptomically defined cell types
Transcriptomic studies of the brain show a hierarchical cell type organization. Similar cell types also show differences between different brain regions. The hierarchy reflects the evolutionary cell origins and relationships. Comparisons between mouse and human cell type trees show that they are conserved across species.
This data also shows that there are continuous and discrete variations between cell types. The discrete differences tend to be at the top of the hierarchies and the continuous one on the lower branches. While cells on different transcriptomic profiles between ends of the tree, the variation between them is a continuum which can make it difficult to divide the cells into types and get exact type counts.
Correspondence between transcriptomic cell types and other cellular properties
Cell type definitions become useful when they are associated with cell function. The cell type taxonomy needs to be linked to anatomical and functional data, to evaluate the transcriptomic classification. Data on other cellular properties can be used to clean the transcriptomic data and pull out the molecular features involved in classification.
Cell types versus cell states
One of the main questions in classifying cell types is at what point is one cell type really just another state of another cell type? This question is particularly relevant at the lower branches of the cell type hierarchies where the differences between types become increasingly continuous. It can be difficult to delineate between types and states during development and between individual and environmental variations.
In the brain, non-neuronal cell types have a number of different cell states under different environmental and disease conditions. These cell types can change their function and activity and transcriptome.
While single-cell transcriptomic data is usually a snapshot of a cell in time, comparing transcriptomic data measured over time could reveal how clusters and cell types move in their transcriptomic space. Generally we could assume that transcriptomic changes are more abrupt and varied in the transitions between types than between states.
Cell type development
Cell type development can involve complex trajectories and multiple differentiation steps. There are transcriptional and epigenetic regulatory programs that drive the proliferation and differentiation processes. The regulatory elements may be captured through a combination of epigenomic and spatial transcriptomic profiling with temporal resolution and lineage tracing.
How to define cell type
Cell types are evolutionary products and basic functional units in multicellular organisms. Their definition can start with the use of single-cell transcriptomics-based taxonomy to capture the genetic signatures driving cell-type differentiation. This information can be related to epigenetic regulatory systems and validated with anatomical and functional studies and used to determine how granular definitions should be. The definition process will also need a systematic study of the developmental process to find causal relationships and their molecular features. All this information and analysis will also need a framework like a ‘periodic table of cell types’ to hold them and standardized criteria, nomenclature, tools to enable sharing and reproducible research.