What are the two methods of cell analysis?
scRNAseq data, once ‘cleaned up,’ needs to be analyzed to make any contributions to future research. There are many different analyses that can be performed, but I will discuss two distinct types that have produced significant discoveries: the identification of subpopulations and cell types and pseudo time analysis. One analysis that is commonly used on scRNAseq data, beyond simple frequency quantification, is identifying subpopulation. Cells are groups together based on their expression levels, specifically using principal component analysis or hierarchical clustering. Hierarchical clustering is a method to cluster cells by similarity, where the similarity between the cells within clusters is higher than the similarity that would be expected by chance. Principal component analysis is a somewhat more complicated method that helps to visualize and quantify many different dimensions of variables.
These two methods are both types of subpopulation and cell type analyses, which allow for visualization, quantification, and identification of cell types and sample subpopulations. Since one of the key advances that scRNAseq allows is the illumination of heterogeneity within populations, clustering is a vital part of the analysis. It allows for the actual separation and identification of this heterogeneity into meaningful conclusions. These include identification of new markers (or genes that, if expressed in a certain way, identify a certain cell type) which can be seen in the recent study where radial glia had certain markers during certain moments of their cell-cycle that were previously unknown, i.e., outer subventricular radial glial cells were characterized by ‘genes related to cellular migratory behavior and extracellular matrices, such as HOPX and TNC,’ but ventricular radial glial cells ‘express CRYAB, PDGFD,TAGLN2, FBXO32 and PALLD’. All this to say that due to scRNAseq, the markers to differentiate oRG and vRG are now known.