Library invisible on gel, despite Nanodrop results. Gel extraction problem?

[thdybwf]

I'm having some difficulty with my very first Illumina library prep, and was hoping someone might point out any obvious errors I'm committing in my ignorance. I'm using a protocol that involves size selection via agarose gel, then adapter ligation, then PCR amplification. Everything look a-OK when things are run out on the gel, stained with GelRed. There's a nice, bright smear from which I cut a band, and purify it using the Qiagen MinElute gel extraction kit. I ligate on the adapter and do the PCR. I run it out on a gel to check if it worked and... nada. The ladder is crisp and clear but the PCR product and template are invisible. Where did the DNA go?

OK, so I'm messing something up, probably with the gel extraction if I had to guess. But weirdly, when I analyze the template and PCR product with a Nanodrop, I get the following:

Template:
Nucl. Acid Concentration: 11.7 ng/ul
A260: 0.235
A280: 0.075
260/280: 3.15
260/230: 3.09

PCR:
Nucl. Acid Concentration: 720.2 ng/ul
A260: 14.405
A280: 9.906
260/280: 1.45
260/230: 0.61

There seems to be a decent amount of DNA in both, and it's much greater in the PCR. Does that mean the PCR is working, despite my not being able to see it on the gel? Is some contaminant from the gel purification to blame for these seemingly high DNA concentrations?

I definitely messed a few things up, such as using anhydrous ethanol rather than molecular grade, and not letting the MinElute columns equilibriate to room temperature, and not doing enough washes during the gel extraction. But what do you make of the Nanodrop results?

[nitut]

Hello,
Can you please check if there is an gel extraction to be carried out before Adapter ligation?
If I am not wrong after the Adapter ligation, the protocols says to do AMPure XP Bead purification and then gel extraction.

[ulz_peter]

The high DNA concentration after PCR might also be caused by dNTPs only, since Nanodrop can't distinguish between unbound nucleotides and a PCR product.

[ingrid@illumina]

Your protocol should look like this:
gDNA -> fragment -> repair ends -> adenylate 3' ends -> ligate adapters -> clean-up using AMPure XP beads -> size selection (gel extraction) -> PCR amplification -> validation -> normalization & pooling

Try using Picogreen or Qubit to validate your starting material and a BioAnalyzer for the best validation of your final libraries.

[luc]

thdybwf,

your PCR Nanodrop data look fishy (for example the 260/280 ratio). Likely, your PCR did not work (because the ligation did not work, because ... ?).
Please post the detailed protocol that you were following. How much DNA did you start out with?

[pmiguel]

But weirdly, when I analyze the template and PCR product with a Nanodrop, I get the following:

Template:
Nucl. Acid Concentration: 11.7 ng/ul
A260: 0.235
A280: 0.075
260/280: 3.15
260/230: 3.09

PCR:
Nucl. Acid Concentration: 720.2 ng/ul
A260: 14.405
A280: 9.906
260/280: 1.45
260/230: 0.61

There seems to be a decent amount of DNA in both, and it's much greater in the PCR.

UV spectrophotometry is often misleading. For the reasons mentioned by others in this thread and a bunch of other ones, including some I detail here.

The nanodrop is easy to use, and the UV spectrum it plots can give you useful information -- if you bother to look at it. But whenever you perform an assay, you want to know its limits of sensitivity as well as anything that might confound it reading only (or mostly) what you are interested in (DNA concentration).

ingrid@illumina gives good about as good advice as can be packed into a couple of sentences. However Illumina (along with many other companies) does its level best to "black box" its protocols. That is, they tend to hide salient details in a misguided attempt to prevent others from reverse-engineering their protocols. That puts them firmly in the "part of the problem" camp in my opinion.

But, hey, I can't entirely blame Illumina because nobody seems to have taught you the basics of what a UV spectrophotometer detects. To a first approximation the answer is "lots of stuff", including DNA and RNA, but also nucleotides, acetic acid or salts thereof, salts of guanidium, phenol, DTT, etc. etc. etc. Don't feel too bad about it personally, more than 90% of bench scientists in molecular biology I meet are in the same boat. (Well, I am at a University, so these scientists are largely students...)

Then there is the issue of sensitivity. Can you detect a 10 nM library on a nanodrop? Worthwhile to keep a couple of "rules of thumb" in mind as you navigate a protocol. Some details here. But to a first approximation:

1 ul of a 1 nM solution contains 1 billion molecules (2nd approx.: ~60% of that.)

1 ng of a 1 kbp fragment is 1 billion molecules (2nd approx.: ~93% of that.)


Therefore if your library is 10 nM, that would be about 6 billion molecules/ul. If these molecules are 1 kbp, then that is about 6 ng/ul. But they are probably closer to 1/2 that length. So 3 ng/ul.

Can you reliably detect DNA at 3 ng/ul on a nanodrop? Not on ours. Below 10 ng/ul, things get pretty sketchy. How about on an agarose gel? Too many variables to say for sure, but probably not unless your library is in a very tight size range or you have a very sensitive system set up.

Hence the wisdom embedded in "ingrid@illuminas" advice -- go with an Agilent Bioanalyzer to QC your library. Using a high sensitivity chip you are good down into the pg range. Or fluorimetry/qPCR -- except you don't see the size of your products that way.

--
Phillip