It is my understanding that neurons can have upwards of 10,000 synapses and at many of those synapses, they may have a mitochondrion that is actively transcribing genes. As I understand it, scRNA-seq assesses the expression of genes by matching RNA transcripts to barcoded tags that serve as proxies for the expression of the genes the RNA transcripts came from. When people do scRNA-seq on neurons, do they often take into account what I presume is a much higher level of mitochondrial gene expression than in other sorts of cells? My main area of knowledge is brain cancer and I don't recall having seen this explicitly taken into account in papers I've read that look at brain tumors' interactions with neurons. Moreover, I know that when people do scRNA-seq analysis, it's common practice to filter out cells whose transcriptomes have more than 5% mitochondrial gene expression.
In addition, if the ratio of barcodes-to-gene transcripts differs between genes (for example, if the ratio of EGFR-barcode to EGFR is 10:1 but the ratio of CD44-barcode to CD44 is 1:1), it seems to me that these different ratios could skew the recorded relative expression of these genes. How do scRNA-seq platforms control for this?
