Why is the fraction of mRNA transcripts captured per cell so low?

Question

[jonas.andersson]

Hi!

I'm analysing scRNA-seq data from 10x genomics chromium 3' v3. What I have learned is that the technique captures ~30-32% of the total mRNA transcripts in each cell.

10x writes on the FAQ page: "Please note that detection of a transcript is a stochastic event, and the majority of the transcriptome is captured as a result of sequencing multiple single cells from the same population."

I read somewhere, I don't remember exactly where but it was some paper about scRNAseq, where they stated that the limitations in the capture rate is because of the reverse transcriptase and how efficient it is to sample mRNA from each cell.

The question is why? why can't the RT manage to capture 100% or close to at least? Is it possible to improve the rate more?

Thankful if someone could please explain this to me:)

best
Jonas

[GenoMax]

Is it possible to improve the rate more?

Some questions have no easy/complete answers. It may simply be the limitation of the current "state of the art" in technology. This efficiency is bound to improve over time but I doubt if we will realistically get to 100%.

Answer 1

[LChart]

It's worth pointing out that RT isn't responsible for capturing the RNA per-se, it's priming to oligo-dTs on the capture bead itself (or, more recently, to rRNA-depleted random hexamers). As mentioned, rRNA will nonetheless bind to these sequences as they contain both polyA sequences and hexamers which, though more common in mRNA, still can occur in rRNA.

Fundamentally, 100% capture is strictly impossible, as there is both a rate of association onto the capture sequence and a rate of dissociation from the capture sequence; though it should be possible to get arbitrarily close.

In addition, length of the polyA tail impacts this efficiency; there are likely crowding conditions on the bead itself, and the speed of the reaction will depend on cDNA concentration.

My guess is that, within a single droplet, the RNA concentration is low enough that to get a capture rate of 50% or 60%, you may have to incubate the bead and cDNA for a prohibitively long time.

It's precisely this low overall efficiency, which means that only a fraction of primers on the bead are ever used in a single droplet, that enables recent barcode-and-overload approaches (https://www.nature.com/articles/s41592-021-01153-z) which significantly improve per-bead efficiency with little drop in per-cell efficiency.

[jonas.andersson]

Thx for the answer, much appreciated!