[SunPenguin]

Can you describe your protocol with more detail, including what your DNA (or cDNA? you mentioned a transcript) sample consists of and what you're amplifying? I've never had to resort to repeated rounds of purification and PCR -- there might be a simpler way to do this by e.g. touchdown PCR.

jdk787's suggestion of a denaturing gel will definitely clear things up. You also might want to run your gDNA/cDNA sample out to see if you're getting high-molecular weight products that are beyond the range of the Bioanalyzer chip you're running. Alternatively, if you have access to a Tapestation the gDNA Screentapes are quite decent at resolving larger fragment sizes.

It goes through multiple rounds because each round uses a different set of primers that are nested to each other. I start with a 5'Race RT (no repeat, just 1 cycle), and then two rounds of nested PCR, then 1 final round of PCR to attach Illumina seqeuencing handles (as opposed to ligation, which in my hands hadn't been efficient).

Right now I've definitely not optimized all the parameters. Library prep is definitely still a bit difficult for me to grasp.

Hmm I think it definitely could be overamplification, since it does seem on the samples that I start with lower concentration of template (just by checking gel band intensity) I tend to have less of this kind of issue. The qPCR results for these sample wasn't super high though, so it's a little hard for me to imagine the reaction being depleted of primers (it's added to the reaction at 0.5uM). Maybe I'm somehow losing amplified product to annealing due to over cycling?

I'm also using Phusion hot start for this reaction. I'm wondering maybe that has something to do with it? perhaps not enough Mg?

[jdk787]

The qPCR results for these sample wasn't super high though, so it's a little hard for me to imagine the reaction being depleted of primers (it's added to the reaction at 0.5uM). Maybe I'm somehow losing amplified product to annealing due to over cycling?

What is the molarity of your final product? Your BA traces look like your libraries are pretty concentrated?

Also, is the final concentration of primers in your reaction 0.5uM or are they 0.5uM before you add them? If they are 0.5uM in the reaction (2uL of 12.5um in 50 uL) I think you should be ok depending on how much material you have going in to the PCR reaction.

[jwfoley]

The graph is consistent with my experience; 0.4X isn't going to work and even 0.5X is dodgy. For the record, you really should recalibrate your SPRI conditions for your specific reaction mix, because certain components (like magnesium, or especially PEG from a "quick ligation") change the binding chemistry.

But then your Bioanalyzer trace looks perfect except for being overamplified.

Instead of a denaturing gel, you can also denature a sample yourself (a couple of minutes at 95 C) and then run it on a Bioanalyzer RNA chip to profile the ssDNA. This won't get artifacts from daisy chains, bubbles, or whatever overamplification causes.

Also FYI Phusion has been obsolete for some time now; NEB uses the new Q5 polymerase in its library kits, and all the cool kids are using Kapa HiFi, even for amplicons in the 10+ kb range.

Titrate that PCR. Just make a bunch of duplicate tubes and take one out after each cycle, then purify per usual and run on the Bioanalyzer (DNA chip). I bet you'll see that a few cycles lower gives you just as big a peak in the desired range and none of the long tail. FYI overamplified libraries sequence just fine; the only problem is that you can't believe the Bioanalyzer numbers to quantify them, but qPCR doesn't care.

[jwfoley]

The total number of cycles right now sit at 70 (it's a low abundance transcript, so it takes a bit to get enough products).

Is this a typo? Even starting from a single molecule, 70 PCR cycles would give you about 21 g of DNA, according to a quick back-of-the-envelope calculation, though of course you'll exhaust the reagents first (and primers are usually the first thing to go, by design). I'll grant you some loss in the intermediate steps, but still. Overamplification definitely sounds like a possibility.

[SunPenguin]

What is the molarity of your final product? Your BA traces look like your libraries are pretty concentrated?

Also, is the final concentration of primers in your reaction 0.5uM or are they 0.5uM before you add them? If they are 0.5uM in the reaction (2uL of 12.5um in 50 uL) I think you should be ok depending on how much material you have going in to the PCR reaction.

The final concentration of the primers are 0.5uM. BioA estimates about 5-10nM on these peaks, and qPCR estimates about 10-20nM. I don't think that's super concentrated?

Edit: actually I looked at the traces again. The ones that show more of a shoulder have 5-10nM estimated by bioA. The ones that have less of a shoulder shows upward of 50-60 nM by bioA.

[SunPenguin]

Is this a typo? Even starting from a single molecule, 70 PCR cycles would give you about 21 g of DNA, according to a quick back-of-the-envelope calculation, though of course you'll exhaust the reagents first (and primers are usually the first thing to go, by design). I'll grant you some loss in the intermediate steps, but still. Overamplification definitely sounds like a possibility.

hmm no it's not a typo, unfortunately. We started with very high cycle numbers and more rounds of PCR, in an attempt to make sure we're getting the right amplification. 70 cycles is divided in 3 steps, and only a small aliquot of product from each step is carried into the next. For gene amplification with nested PCR, I thought that's not really that unusual, though it may be a bit old fashion?

Phusion in our hands just gives the best amplification, so we've stuck with it. I've also tried Platinum Taq (in an effort to reduce chimeric product in exchange for higher error rate) and Kapa, though neither of those have worked as well. I haven't attempted to optimized those enzymes, however.

[SunPenguin]

The graph is consistent with my experience; 0.4X isn't going to work and even 0.5X is dodgy. For the record, you really should recalibrate your SPRI conditions for your specific reaction mix, because certain components (like magnesium, or especially PEG from a "quick ligation") change the binding chemistry.

But then your Bioanalyzer trace looks perfect except for being overamplified.

Instead of a denaturing gel, you can also denature a sample yourself (a couple of minutes at 95 C) and then run it on a Bioanalyzer RNA chip to profile the ssDNA. This won't get artifacts from daisy chains, bubbles, or whatever overamplification causes.

I see. I'll definitely look into recalibrating the SPRI.

After you denature the DNA at 95, do you then just let it cool back down to roomtempt? or do you have to try to run it as soon as possible?

[jwfoley]

After you denature the DNA at 95, do you then just let it cool back down to roomtempt? or do you have to try to run it as soon as possible?

I've always just run it immediately, so I don't know what happens if you wait.

[SunPenguin]

I just ran a titrating PCR like you said, and ran a quick E-gel on it. It does seem reducing the template and cycle number concentrate the product more and produce less smear! though even just a few cycles too low, the product band decrease dramatically in intensity.

I'll put it through spri and run it on the BioA and update.