Enrichment PCR Failure

[kbusseq]

Hi all,

For the past few weeks we've been unable to successfully amplify our post-library prep samples. We use the Apollo324 automated system to do the library preps, then do the PCR enrichment and Ampure clean-up by hand. What we've seen in the past is decent yield coming off the Apollo and better yield following PCR (10 cycles for most samples) and clean-up (as one would expect!) But rather abruptly it seems that our PCR amplification has just stopped working. Instead of increased concentration following PCR, we have greatly reduced concentration, to the point where there's not enough sample to load on the HiSeq. I have a couple pictures of Bioanalyzer traces from the tests we've been doing this week, of the same sample before PCR amplification and then again after PCR amplification and clean-up, so you can see what I mean; we've seen samples worse than this as well.

One interesting aspect is that if we take the post-amplification sample and run another 10 cycles of PCR, we see concentrations more like what we would have expected to see after the first round of amplification (I don't have a bioanalyzer trace file for this, but we confirmed it with the nanodrop and an agarose gel). So apparently something is working with the ligation and the amplification - maybe just not very efficiently or consistently? Do you have any ideas on what could be causing the concentration drop after that first round of amplification?

(I did notice a similar thread from a couple years ago when I was searching the forums for ideas, but there was no resolution to the problem, so I figured it would be best to just start a new thread with my question. Hopefully that is ok!)

Thanks so much,

Kristina

[kwaraska]

Just some thoughts---

What beads are you using for clean up on the robot? Are they expired or close to it?
Is this a new batch of enzyme for the PCR?

Do you have the most updated protocols for the Apollo?...I feel like they are constantly upgrading and maybe your program has residual ethanol in the sample. That would cause PCR to fail the first time but once cleaned up it would be okay.


I don't know if these are the answer but at least it is something to look at.

[kbusseq]

We're using Ampure beads for clean up; they have about a year left before they're supposed to expire. We tried a number of different enzymes for the PCR (some old, some new) and saw no difference in the performance. We do have the most updated protocol for the Apollo, and the rep actually came out here a couple weeks ago to do a PM and a test run, so I'm doubtful that it's a machine issue.

The suggestion about residual ethanol is definitely something to look at, I agree. It wouldn't be that hard to control for the possibility, we'd just have to speed vac and rehydrate the samples. And at this point I'm up for trying pretty much anything in hopes of identifying the problem, and this would be a pretty simple step to add.

Thanks for the thoughts!

[susma]

Hi guys,

I am so sorry that I bring up my problem here. I have already posted my problem in the forum, but nobody answered me You seem to have experience in this area, would you please help me?
This is my problem:


It is my first time that I make DNA libraries for RNA sequencing. After using the ScriptSeq V2 library preparation kit I ended up with the attached results. I have run one sample with different concentrations on the High sensitivity DNA chip. Actually I do not have any idea if these peaks are OK or not.

I would really appreciate if someone gives me an advise or explanation, I am running out of time for this project!

[kbusseq]

Hi Susma,

The peaks don't look that great. Ideally, even for RNA libraries, you want to see a smoother curve than what you've got here. What I've noticed in my experience, though, is that the starting quality of the RNA really influences what the final library curve looks like. Do you know the RIN score for your starting RNA? We've sequenced some samples that had a starting score around 8-9 and seen peaks as smooth as a genomic DNA peak, and we've sequenced other samples that had starting scores around 3-4 and seen peaks that are more similar to what you have here (slightly flatter and more jagged).

I haven't done as much on the informatics side, but from what I've heard you're not going to get the best quality data out of samples that look like this. The presence of all the smaller jagged peaks within the Bioanalyzer curve suggests that certain RNA transcript populations are over-represented in comparison to other transcript populations (assuming you're primarily working with mRNA here), so you probably won't see the full distribution of transcripts that you're hoping for. On the other hand, it may be the best you can do (for instance, if the RNA is extracted from FFPE samples, you might not be able to get any higher quality than this). So I can't really tell you whether or not your samples are ok or not - it depends on a lot of different variables and in the end you have to make that call for yourself.

Good luck!

[susma]

Thank you very much for your time.

The fact is that this mRNA used here is from bacteria and about 90% of rRNAs have been removed. RIN is 9.3 and 260/280 and 230/280 ratios are in 1.8-2.0 range.

Do you think it might be the high mRNA input concentration problem, since I do not have Qubit and I just used Nanodrop to measure the concentration?

I have contacted people in Epicentre technical support for help.

Best regards,
Susan

[nettybetty]

Hi Kristina,

I know it's only a few days, but how have you faired with your samples? Have you had a chance to do some troubleshooting? I'm wondering if I have the same issue as you and am rather perplexed about where to go next!

I did an automated TruSeq DNA library prep with 24 samples, and the samples in column 3 of the plate have failed to PCR and the rest look fine. They were a different DNA source but were prepped along the rest so I don't think it's a reagent or bead issue. All QC pre-library looked fine, so it's very perplexing. However I can believe it could be an ethanol issue - perhaps I need to re-teach some of the positions on my robot. I've only just started using an automated protocol, so I kept a very close eye on the robot and have all the reagent plates etc. and nothing was amiss during the run.

I haven't tried doing another PCR as I haven't yet done the post-PCR SPRI - when I saw the low levels of the dirty result I was very hesitant to do a clean up. When you do a second PCR, how are you setting this up? And have you actually put any of these 2 rounds of PCR libraries on the HiSeq? How have they fared?

[microgirl123]

A couple of thoughts. Are you KAPA quantifying your library directly from the Apollo? If the ligation isn't working well (possibly you have a bad batch of ligase or it's been mistreated), then the PCR reaction won't be able to find much template to amplify. We also have an Apollo and (prior to our latest script update) usually saw a lot more DNA on the Bioanalyzer relative to functional library quantification through qPCR. We don't usually do PCR enrichment at all on our Apollo samples because there's always been enough for the MiSeq!

Also, for the bead cleanup, is your ethanol fresh and properly mixed?