Reply
Highlighted

Chromium single-cell 3' RNA-Seq on HiSeq 4000: initial results?

Posted By: james, on Jan 3, 2017 at 7:01 AM

We've just completed our first run on the HiSeq 4000 of V2 single-cell libraries and the results look good. However we're not really sure how good they should look so I'd be interested to get feedback from 10X, and from other users.

 

Number of reads: We got almost 500 million raw reads that were filtered to 400 million so this was a higher number of reads than we might have expected.

Error rates: we saw the error rate on the cDNA read was 0.7% on PhiX.

%Q30: 75% of cDNA bases were >Q30.

I've attached a screenshot from BaseSpace below.

Screen Shot 2017-01-03 at 14.58.39.png

 

 

11 Replies

Re: Chromium single-cell 3' RNA-Seq on HiSeq 4000: initial results?

Posted By: shauna-10x, on Jan 4, 2017 at 10:58 AM

Hi James,

 

Thanks for posting to the 10x Community.  I've forwarded your question to support for review.  In the meantime, this is a great thread for others to share their sequencing results using single cell libraries, too.

 

Best,

Shauna

Re: Chromium single-cell 3' RNA-Seq on HiSeq 4000: initial results?

Posted By: timstuart, on Jan 16, 2017 at 12:13 AM

Hi James, we just completed a run with V2 chemistry on the NextSeq (read1=26bp, 8bp sample index, read2=116bp). Sequencing depth is too low to say much about our experiment, but we noticed some potential problems with the seq run/library.

 

We found read 2 looks a bit odd, with a much higher error rate and lower quality scores.

 

Read1:

% error: 0.12

%>Q30: 98.5

 

Read2:

% error: 2.23

%>Q30: 71.2

 

Attached are some screenshots from basespace. I'm trying to work out if the low quality of read 2 is due to reading into the poly A tails, but it would be helpful to know if others have experienced this issue too. Mean fragment size was ~500 bp so I would be surprised if there was much poly-A tail being sequenced.

 

Screen Shot 2017-01-16 at 4.09.13 pm.pngScreen Shot 2017-01-16 at 4.09.24 pm.pngScreen Shot 2017-01-16 at 4.09.30 pm.png

Re: Chromium single-cell 3' RNA-Seq on HiSeq 4000: initial results?

Posted By: james, on Feb 21, 2017 at 6:56 AM

It certainly looked much worse than ours but I've no idea why. What did 10X say? Have you completed any further runs since this one?

Re: Chromium single-cell 3' RNA-Seq on HiSeq 4000: initial results?

Posted By: mg, on Feb 22, 2017 at 8:03 AM

So far we have sequenced on HiSeq4000 8 different pools of different plexity coming from different projects and sample types. In all cases we got around 400M reads PF.

All lanes were clustered at 240pM, most lanes were run as R1:26bp, R2:98bp.

Our 1st  flow cell is the one mentioned by James at the beginning with around 75%>Q30 of the R2.

2nd flow cell was actually even better with around 80%>Q30 of R2.

We spiked in 1% PhiX to all lanes but we got around 0.5% aligned. Error rates on R2 vary from 0.7 to 1.2.

 

When running PE150 we got around 70%>Q30 of R2, the error rate increased to 1.4 and the quality started to drop around 100th cycle of R2. Reading into polyA in R1 results in extreme drop in quality.

 PE150.jpgWe also see the low diversity in the first 25-30 cycles of R2. Would that be caused by the switch oligo? I don't etirely get how and why in some samples it's more pronounced.

 

Interestingly the second cycle of R2 has %>Q30=0 on all lanes so far. This seems to be specific to this library type on HiSeq4000 machine.

Marta