Smart-seq2 sample prep questions

[PrimerBuffalo]

Hi All,

I have a couple questions about the Smart-seq2 method for amplifying single cells (Sandberg 2014), and I hope someone with experience in this method might have some answers:

1) Why are dNTPs present in the catching buffer that you sort cells into? Couldn't they just be added with the RTase after the initial binding of the OligoDT to the polyA tail?

2) What are the drawbacks of having OligoDTs or dNTPs in excess to what is written in the original protocol? I am worried about the concentrations not being enough for a highly transcriptional cell.


Thanks in advance for the help!

[cmbetts]

1) If I remember right, they mention in the paper that they got a statistically significant yield increase adding dNTPs at that step. It definitely would still work, just with somewhat lower efficiency if it's added w/ RT
2) Too much dT can generate artifacts, more dNTPs probably wouldn't have a noticeable effect. That same protocol can be used without modifying the RT conditions for bulk samples with >10ng RNA (only PCR cycles are reduced), so you don't need to worry about insufficient primer for any single cell applications. There are several orders of magnitude more primer than mRNA in the cell. You should do some preliminary characterization experiments with your cells to dial in the correct number of PCR cycles, so you might end up shaving off a cycle or two from the default if your cells are particularly active.

[Kaizen]

Do you have any experience using Smartseq2 for bulk samples? from 1000-10000 immune cells.
I would like to know what's the maximum number of cells that you can process in a 96 well plate using bead purification without modifiying the RT-PCR and also the maximum number of cycles recommended given a certain number of cells
Thanks

[Simone78]

Do you have any experience using Smartseq2 for bulk samples? from 1000-10000 immune cells.
I would like to know what's the maximum number of cells that you can process in a 96 well plate using bead purification without modifiying the RT-PCR and also the maximum number of cycles recommended given a certain number of cells
Thanks

Hi,
Smart-seq2 works for bulk samples, however 10K cells might be too much. Among the possible drawbacks:
- lysis might be incomplete. To avoid this I would increase the % of Triton in the LB. You can surely increase it to 1-2% (as in the STRT-seq 2i protocol). Higher % of Triton seem also not to harm the RT enzyme but I would be cautious with that. Try to decrease the number of cells instead.
- You might run out of some components later in the pre-ampl reaction. The most likely "candidates" are dNTPs and oligos (all). I would suggest to run a sort of titration experiment, something along the line: if you start from 1K cells do you get more cDNA after pre-ampl when using 2 uM final of TSO vs 1 uM? I know it´s just empirical but so far I didn´t come up with something better, I am afraid. If you don´t want to modify the original protocol I would try to stay in the range 100-500 cells. Maybe it works also with 1K cells, but I would at least increase the % of Triton. While the amount of RNA/cell is known for many cell types (and reported to be around 10 pg tot RNA for cell lines), I couldn´t find reliable numbers for immune cells. probably 0.1-1 pg/cell?

With single T/B cells, ILC or monocytes I do 22-24 PCR cycles for the pre-ampl reaction. With >1 cell I would adjust the number of cycles accordingly. As a guideline: for 100 cells you might find 18-20 cycles sufficient; for 1K cells maybe 16-18, and so on. In the end you just want to have enough cDNA to be able to see it on the Agilent BA or TS. For tagmentation, both with the kit or the home-made Tn5 you would anyway use 1 ng cDNA or less. I would elute in smaller volumes after bead cleanup rather than increasing the number of pre-ampl cycles to be able to see it on the BA/TS.

I hope this helps!
/Simone