RNA-seq from single cells?

[ChristmasSunflower]

Originally Posted by cmbetts

There's tons of low MW products there indicating massive RNA degradation, which you have in both your cell and purified RNA samples. Did you check the RIN of your pure RNA? If it's high, then you've got an RNase contamination problem, otherwise you need to look at your cell/RNA isolation procedure. I would not proceed to library prep and sequencing with those samples.

Yes, RNA control quality is checked by Bioanalyzer with the perfect RIN=10. I did parallel 2 technical replicates with RNA 1ng, RNA100pg and RNA10pg and just didn't get chance to load to this 11-well HS chip. But if you looked at my negative control, there is also big peak in the left size so I initially thought it's some other things.

Thanks

[ChristmasSunflower]

Yes, RNA control quality is checked by Bioanalyzer with the perfect RIN=10. I did parallel 2 technical replicates with RNA 1ng, RNA100pg and RNA10pg and just didn't get chance to load to this 11-well HS chip. But if you looked at my negative control, there is also big peak in the left size so I initially thought it's some other things.

Thanks

[ChristmasSunflower]

Originally Posted by Simone78

I agree, the RNA controls should be much better and give tons of cDNA, especially the 100 pg and 1 ng. Lots of degraded stuff or leftover primers in general. S2 and S7 are decent, though. S3 and S4 still have cDNA left, the others are all degraded. You might want to be more stringent with bead cleanup (0.8X is sufficient) and/or decrease the amount of oligos you are using. Which cells are these, by the way? That would determine the number of PCR cycles and help understanding whether it´s possible to get some good cDNA in general. Although Smart-seq2 is rather robust, not all cells give good results and you might need to play around with the lysis buffer.

They are mouse sensory neuron cells. I wonder why there are same things in the left side in my negative controls. I thought I got some cDNA amplifications in experimental samples. Yes, these pure RNA control results are puzzled me.

Thanks

[Simone78]

Originally Posted by ChristmasSunflower

They are mouse sensory neuron cells. I wonder why there are same things in the left side in my negative controls. I thought I got some cDNA amplifications in experimental samples. Yes, these pure RNA control results are puzzled me.

Thanks

the peak on the left, in my opinion, is primers/dimers. Unused primers, either because there is no RNA (NC) or degraded RNA will be amplified and accumulate in the reaction. If the RNA is little or degraded, the amount of primers is in large excess (it´s in large excess anyway, but this is another story...). Maybe the amount is so high that the bead cleanup can´t remove them entirely. Cutoff of beads is around 100 bp but even shorter fragments are retained.

[ChristmasSunflower]

Originally Posted by Simone78

the peak on the left, in my opinion, is primers/dimers. Unused primers, either because there is no RNA (NC) or degraded RNA will be amplified and accumulate in the reaction. If the RNA is little or degraded, the amount of primers is in large excess (it´s in large excess anyway, but this is another story...). Maybe the amount is so high that the bead cleanup can´t remove them entirely. Cutoff of beads is around 100 bp but even shorter fragments are retained.

Hi Simone,

I thought they are some primer leftovers after beads cleanup. But I do see the amplification of cDNA in most experimental samples, such as S#1, , 3, 4 5 and 7, right? There are consistent peak around 2Kb in most of experimental samples, I thought it's a good sign.

Thanks

[sbarberan]

Talking about which cells are used, is there anyone here with experience with single-cell experiments with MCF7?

Also, Simeone78, when you say play around with the lysis buffer, what exactly do you play around with?

[ChristmasSunflower]

Originally Posted by sbarberan

Talking about which cells are used, is there anyone here with experience with single-cell experiments with MCF7?

Also, Simeone78, when you say play around with the lysis buffer, what exactly do you play around with?

Yes, I want to know how to play with lysis buffer?

Thanks

[Simone78]

Originally Posted by ChristmasSunflower

Yes, I want to know how to play with lysis buffer?

Thanks

One could try to increase the amount of Triton or use other (non-ionic) detergents such as Tween-20, Igepal CA-630, NP-40 or BSA (as in Svec et al. 2013, PMID: 24224157). Maybe some other RNAse inhibitor would help. In my opinion, SuperaseIn (Thermo Fisher) is better than Rec RNAse inh. from Takara used for the Smart-seq2 protocol. We just don´t use SuperaseIn because of the higher cost.
If you use the RNAse inh. from Takara you might want to add some DTT (as in the STRT-seq and STRT-seq 2i papers from S. Linnarsson), since the RNAse inh. requires DTT for working in an optimal way.
Definitely don´t use SDS, as ionic detergents kill the enzymes, no matter which concentration you do the lysis.
In a recent paper I have also seen RLT buffer (5M guanidine isothiocyanate) but I can´t recall which one, sorry. It might be some kind of modified scATAC-seq paper.
There are also many patents exploring the subject. I am especially fond of US 2011/0136180A1, for example

Best,
Simone

[ChristmasSunflower]

Thanks, Simone! You have a big play in this discussion topic and I believe many people have benefited from this discussion. Really appreciate!


Thanks

[thang198]

Originally Posted by Simone78

we have recently published a paper where we describe Smart-seq2, an improvement of the Smarter kit. Check it out! You don´need any kit to make your cDNA and you´re going to save a lot of money!

http://www.ncbi.nlm.nih.gov/pubmed/24385147
http://www.ncbi.nlm.nih.gov/pubmed/24056875

Hi Simone78,

Would you please let me know that I can modify your Nature protocol paper "Full-length RNA-seq from single cells using Smart-seq2" for the bulk RNA samples?
I do have 60 different RNA samples but the Truseq from Illumina is expensive. I found that your protocol may reduce the cost since the Nextera is cheaper.
Thank you,

[Simone78]

Hi,
in principle you could use the protocol for single cell as it is, provided that the input RNA is not too high. I don´t know exactly, maybe up to 1 ng. If you start form several ng you should increase the amount of some reagents, to avoid their depletion during the RT or PCR. I would (as an example) double the amount of TSO, ISPCR and oligodT primers and increase also the conc of dNTPs. Again, you should run some kind of titration to see what reagent becomes the limiting factor in your reaction. If planning to sort hundreds of cells in each well, you might also consider using a higher conc of Triton to ensure proper lysis (or a stronger buffer like guadine thiocyanate/hydrochloride, RLT, etc). Concentration of Triton of 1-2% do NOT have a negative effect on the RT reaction. If you need extra info just give me more details about your experiments

[Yin]

Hi all,

I'm new to single cell RNAseq and just started to worked on SMARTseq2,
I followed the nature protocol (PMID: 24385147), but one tricky thing is that we don't have enough room to get a individual hood for single cell experiments, so I have to use shared hood with others for RNA operation. For each sample, 10 colorectal cancer cell HT29 cells were picked up with micro manipulator then released into a PCR tube with lysis buffer in tube. After that samples were incubated at 75C for 5min, then followed the SMARTseq2 protocol described in paper, 10 cycles in transcription and 17 cycles in preamplification, but the transcriptase I used was Maxima H Minus Reverse Transcriptase (200 U/µL) (EP0751, Thermo).
Actually, I got some results that were not too bad at the beginning, but in current few weeks I suffered from severe RNA degradation issue, although the cDNA yield could reach ~80ng, the peak looked really wired, even I doubled the RNase inhibitor, there's no improvement. Is it too harsh to do SMARTseq with my conditions? Hope you experts could help me and give me some suggestions. Thank you!
Attached is the electropherogram from analyzer https://1drv.ms/u/s!AmIgzq2vDK6fliMiOkXyIEPnfozZ

[lijing2019]

I am trying to do single cell RNAseq for human T cells using Smartseq2. I used the smartscribe to do RT and KAPA HiFi to do pre-amplification. I followed the protocol from my colleague, which leaves out 1M Betaine and 6mM MgCl2 for the first 5 samples. Later I heard from another colleague that adding Betaine will increase the cDNA yield so I added these two reagents for the rest of my samples. Other settings and reagents were kept the same. When I did the QC for the samples by fragment analyzer after bead clean-up, I found with Betaine and MgCl2 the average concentration of the preamplified cDNA decreased dramatically, from 0.5ng/uL to 0.03 ng/uL. So I am wondering what could be the problem.

[Simone78]

Originally Posted by lijing2019

I am trying to do single cell RNAseq for human T cells using Smartseq2. I used the smartscribe to do RT and KAPA HiFi to do pre-amplification. I followed the protocol from my colleague, which leaves out 1M Betaine and 6mM MgCl2 for the first 5 samples. Later I heard from another colleague that adding Betaine will increase the cDNA yield so I added these two reagents for the rest of my samples. Other settings and reagents were kept the same. When I did the QC for the samples by fragment analyzer after bead clean-up, I found with Betaine and MgCl2 the average concentration of the preamplified cDNA decreased dramatically, from 0.5ng/uL to 0.03 ng/uL. So I am wondering what could be the problem.

That´s really strange, it is generally the opposite: betaine and MgCl2 gives better yield, even when only just one of the two is included. Could you post the Bioanalyzer plots to get an idea of how the cDNA looks like? Although for you I don´t think it´s the case, it is good practice to include a 10 pg total RNA sample as a sort of control, in order to avoid the cell-to-cell variability.
You are talking about a small number of samples, how did you pick them? By hand? How long the picking take might also have an impact on the RNA quality.
Best,
Simone

[lijing2019]

Hi Simone,
I have 4 patients without adding Betaine and MgCl2, the final yields of preamplified cDNA is almost the same, about 0.5ng/uLx25uL. 6 patients with Betaine and MgCl2, one of the patient seemed to have RNA degradation (could you check whether it is RNA degradation?), the other patients all have low concentration, about 0.03ng/uLx25uL. I attached the representative fragment analysis results. Each pateint has 400 cells and I QC 100 cells from each of them, the results are consistent for each protocol.
I usually thawed the cells and rested them overnight, then FACS sorted them into lysis buffer. I run 22 cycles at the preamplification step. The picking takes about 3 minutes for each plate.
I am going to try sequence some of them anyway. Do you think the patient which seemed to have RNA degradation will work?

[lijing2019]

Originally Posted by Simone78

That´s really strange, it is generally the opposite: betaine and MgCl2 gives better yield, even when only just one of the two is included. Could you post the Bioanalyzer plots to get an idea of how the cDNA looks like? Although for you I don´t think it´s the case, it is good practice to include a 10 pg total RNA sample as a sort of control, in order to avoid the cell-to-cell variability.
You are talking about a small number of samples, how did you pick them? By hand? How long the picking take might also have an impact on the RNA quality.
Best,
Simone

Hi Simone,
I have 4 patients without adding Betaine and MgCl2, the final yields of preamplified cDNA were almost the same, about 0.5ng/uLx25uL. 6 patients with Betaine and MgCl2, one of the patient seemed to have RNA degradation (could you check whether it is RNA degradation?), the other 5 patients all have low concentration, about 0.03ng/uLx25uL. I attached the representative fragment analysis results, ERCC were added in every well. Each pateint has 400 cells and I QC 100 cells from each of them, the results are consistent for each protocol.
I usually thawed the cells and rested them overnight, then FACS sorted them into lysis buffer. I run 22 cycles at the preamplification step. The picking takes about 3 minutes for each plate.
I am going to try sequence some of them anyway. Do you think the patient which seemed to have RNA degradation will work?

[Simone78]

Originally Posted by lijing2019

Hi Simone,
I have 4 patients without adding Betaine and MgCl2, the final yields of preamplified cDNA were almost the same, about 0.5ng/uLx25uL. 6 patients with Betaine and MgCl2, one of the patient seemed to have RNA degradation (could you check whether it is RNA degradation?), the other 5 patients all have low concentration, about 0.03ng/uLx25uL. I attached the representative fragment analysis results, ERCC were added in every well. Each pateint has 400 cells and I QC 100 cells from each of them, the results are consistent for each protocol.
I usually thawed the cells and rested them overnight, then FACS sorted them into lysis buffer. I run 22 cycles at the preamplification step. The picking takes about 3 minutes for each plate.
I am going to try sequence some of them anyway. Do you think the patient which seemed to have RNA degradation will work?

From what I can see you seem to do everything correctly. You might get something out of the degraded cDNA (yes, it seems degraded to me) but you will also have a very strong 3´bias. The Tn5 transposase will tagment anything <40 bp, cDNA as well as primer dimers.
Best,
Simone

[zebrafish_juju]

Hello Simone and all,
Thank you for the deep thought and informative discussion. I learned a lot. I have a question regarding the oligos for the reactions. It looks like the consensus now is to order biotin blocked oligos. Can I still use standard desalt or I have to order HPLC purified ones?

[Simone78]

Originally Posted by zebrafish_juju

Hello Simone and all,
Thank you for the deep thought and informative discussion. I learned a lot. I have a question regarding the oligos for the reactions. It looks like the consensus now is to order biotin blocked oligos. Can I still use standard desalt or I have to order HPLC purified ones?

I personally prefer HPLC-purified oligos when working with single cells but desalted would work as well.
Best,
Simone

[ahmadali]

RNA-Seq, also called RNA sequencing, is a particular technology-based sequencing technique which uses next-generation sequencing to reveal the presence and quantity of RNA in a biological sample at a given moment, analyzing the continuously changing cellular transcriptome.