I noticed that for single-cell mRNA seq, usually N6-12 UMI. But not yet found any paper fully describe how these numbers are calculated. I know roughly a mammalian cell usually has 360,000 mRNA copys of 12,000 genes. But what is roughly the copy number of highest transcipted gene (usually α-actin I assumed)? what is roughly the copy number of GAPDH (but actually different tissue/cell varies quite much right?)?

And why for a typical PE150 mRNA seq (6G/sample), a 8 digit UMI (Don't know why many companies using the TRU-seq rather than SMART-seq library construction for mRNA) is used for a sample of 500ng total mRNA? If self-constructing a library, for 5ug total mRNA, how long of the UMI is enough? just 10 times of the pool and 2 more digits?

Thank you!