[scotto]

Thanks, Phillip. We submitted our troubleshooting form and gsSupportTool output on Thursday and have only gotten an automated email confirming receipt. I don't know how long it usually takes to get a response.

[pmiguel]

Hi scotto,
You probably should follow up. GS-Support is often overwhelmed. If you don't inquire after a couple of days, then you may never get a response.

--
Phillip

[scotto]

Using the shotgun analysis instead of amplicon got most of the parameters in spec. The raw reads are still low, but that's another issue.

[LFMar]

Hi everybody, I just performed an amplicon run to sequence 16S RNA using fusion primers and the "one way read approach". The amplicons have ~700bp and the bioanalyzer profiles were perfect with no primer-dimer peak. Guess what? I obtained a high % of short read/short primer, reads with low quality. The processing was made using amplicon and shotgun pipelines giving the same poor result. Amplicon sequencing using 454 shouldn't be allowed. I would appreciate any advise from you. Thanks in advance.

[pmiguel]

Hi everybody, I just performed an amplicon run to sequence 16S RNA using fusion primers and the "one way read approach". The amplicons have ~700bp and the bioanalyzer profiles were perfect with no primer-dimer peak. Guess what? I obtained a high % of short read/short primer, reads with low quality. The processing was made using amplicon and shotgun pipelines giving the same poor result. Amplicon sequencing using 454 shouldn't be allowed. I would appreciate any advise from you. Thanks in advance.

Okay here is my advice. Take an aliquot of your amplicons. Strand denature them (95 oC 2 minutes) just prior to loading them on an RNA bioanalyzer chip. (Either RNA nano or RNA pico). Often we see confusing results (eg, two peaks for a single amplicon) for our library molecules. Just ignore them. What is important is if you see anything in 150 nt or less region. If so, calculate the molar concentration of those small products. If they are more than 10% or so of your total molar concentration of sample then do 2 cycles of AmpureXP clean up on them. I would not use more than 0.7 vol of AmpureXP to 1 volume of sample for 700 bp amplicons. Then do another RNA bioanalyzer chip to make sure the primer dimers are gone.

The trick here is that small amounts of primer dimer anneal to the 700 bp strands during PCR. PCR clean-ups get rid of short products, but not the short products annealed to long products. Each time you do a small-product removal cycle, some more of the short products denature and re-anneal to other short products and can be removed via size selection. Or at least that is my interpretation of what is happening.

--
Phillip

[TonyBrooks]

We routinely do 16s runs and they almost always look good. I agree with Philip that primer-dimer seems to be a major problem in all 454 amplicon work. All our amplicon libraries undergo gel-cut post pooling. The one library where we didn't do this gave us problems.
FYI I've attached results from our typical run. We use software version 2.5.3 with everything set as default. You have to remember that the Bioanalyser is a plot of DNA mass against size. What's important is not mass of DNA but molarity, so small peaks that are low in length are actually a big problem as they contain a high number of molecules. These are also favoured in emPCR over the longer amplicons you want to sequence. Also I don't think the Bioanalyser is sensitive enough to pick up any low level dimerisation.

It's might also be a good idea to also keep the results of the first NaOH wash during the bead enrichment step. This will contain the reverse strands of the emPCR which you can then run on a gel or Bioanalyser. You will probably need to concentrate on a Qiagen column first after neutralising with a small volume (40ul) of 3M sodium acetate (make sure the Qiagen indicator solution is yellow).

[pmiguel]

We routinely do 16s runs and they almost always look good. I agree with Philip that primer-dimer seems to be a major problem in all 454 amplicon work. All our amplicon libraries undergo gel-cut post pooling. The one library where we didn't do this gave us problems.
FYI I've attached results from our typical run. We use software version 2.5.3 with everything set as default. You have to remember that the Bioanalyser is a plot of DNA mass against size.

Exactly. But with the caveat that single stranded polynucleotides tend have the same mobility or even much lower (larger in appearance) than double stranded polynucleotides of the same length. See here and here for details. Not salient here, but does not seem to be widely known that this is the case for bioanalyzer chips.

What's important is not mass of DNA but molarity, so small peaks that are low in length are actually a big problem as they contain a high number of molecules. These are also favoured in emPCR over the longer amplicons you want to sequence. Also I don't think the Bioanalyser is sensitive enough to pick up any low level dimerisation.

Tony,
It is worse than that. There is nothing to stop a primer dimer strand from annealing to a full length amplicon strand. So even if you get rid of the low molecular weight primer dimers quantitatively your sample may still be contaminated with primer dimers. Denature the strands and they may be there.

--
Phillip

[Anthony.287]

TonyBrooks,
I’ve been wondering if our 16S runs could get better, and then I looked at the thumbnails you posted; ours are performing very similarly. I’ve had trouble in the past getting the enrichment low enough to even bother with sequencing, but between 0.01 and 0.05 cpb seems to be the sweet spot (after quantifying with the KAPA qPCR kit) for 8-20% enrichment.
After following this thread, I’m wondering if it would be feasible to denature the DNA immediately prior to the AMPureXP purification? Will the AMPureXP beads bind ssDNA, or would the RNA SPRI beads be an option? I’ve had some issues with small fragments that don’t show up until sequencing, and AMPure-ing denatured DNA seems like it might alleviate the problem, if it would work.
This is a great thread, and you guys have given some great information!! Keep it up!
Anthony

[MissDNA]

On the enrichment topic, do you guys sequence stuff that enriched over 20%? Most of the times we get enrichment over 20%, we repeat emPCR. Sometimes I wonder if we are being over cautious.

[Anthony.287]

I have sequenced libraries that enriched all the way up to ~60%, due to customer requests and deadlines. Nothing above ~20% has worked well at all. One lab I know of, a beta site for the + upgrade, repeats about half of their emPCR's; anything above 20% they don't bother sequencing.

[MissDNA]

Thanks, Anthony. Good to know we are not over cautious. Also, that most of 454 users seem to have problems with enrichment. We do repeat a emPCR every now and then but certainly nothing close to half. Most of our titration does not perfectly correlate with MV/LV yet we seem to have a good guess-O-meter. It is harder to find the right cpb for MVs.

[ajthomas]

Will the AMPureXP beads bind ssDNA, or would the RNA SPRI beads be an option? I’ve had some issues with small fragments that don’t show up until sequencing, and AMPure-ing denatured DNA seems like it might alleviate the problem, if it would work.

I had a bottle of the RNA SPRI beads that I knew I wasn't going to use before they expired so I called to ask the company if they would work for DNA. I was told that they are the same product, except that the RNA beads have been produced to be sure they are RNAse free and are checked for RNAse contamination, so yes they should work fine for DNA. Assuming that's true, the reverse should also be true--that you should be able to purify ssDNA with AMPure beads. I suspect the ratios will need to be adjusted, but I don't know that for sure.

On the enrichment topic, do you guys sequence stuff that enriched over 20%? Most of the times we get enrichment over 20%, we repeat emPCR. Sometimes I wonder if we are being over cautious.

On my last run, a 8-region plate, I had two samples that were higher than I wanted, but I went ahead and sequenced them anyway. I got 22% enrichment with one amplicon library, but I still got a filter pass rate of ~50%. That was the lowest of the 6 amplicon libraries on the plate, but the highest was only about 55%, so it was right in line with the others. Another library had a 33% enrichment, but I only had about 17% of those reads pass filter. That was a new library that was something new I was trying, however, so I'm not sure how much of that low filter pass rate was due to the high enrichment and how much was due to the library itself.

[MissDNA]

22% is not too far from recommended. The highest we have sequenced was 27% and as far as I remember there was not much difference from the other runs we had performed with the same library. Last week we perform a 4-region 3kbPE run. All the libraries were equally good, 3 of them had enrichment around 5-6% (we decided to repeat emPCR using SV, because originally we had gotten over 20% on MV), the 4th library was at about 14% (they were product of the 1st MV we had done). The 3 libraries with low enrichment had over 70% passed filter, while the one with 14% had only 49%.
I do think that passed filter has a lot to do with the libraries quality, still I am not comfortable to potentially ruin a run by using super enriched beads.
I would not run amplicons that had over 30% enrichment.

[ajthomas]

I normally don't run anything above 20%, but in that case, I had all the others ready to go and couldn't wait any longer. The one that was 22% was a little high, but as you said, it wasn't much above 20% so I went ahead with it. With the other one, I thought the results would probably be not so great, but because of what it was and what I needed out of it, I figured that would be okay. As it turned out, I got the information I needed in spite of the low filter pass rate.

[MissDNA]

Sometimes we have to rush
Usually how much passed filter you get on your amplicon runs?