Problems with input DNA quantification

[skblazer]

Dear all,

I faced the problem of quantification of input DNA for library preparation process. I attached a gel image showing six different DNA preps. According to nanodrop quantification, they should be (from left): 380ng, 940ng, 1080ng, 1030ng, 638ng, 620ng, marker.

But why there are too much difference between the quantification result of nanodrop and EtBr image.

I always resuspend DNA with RNase for 37C for 1.5h, following with Phenol/chloroform treatment and 2X EtOH (1/10 NaAc) precipitation. DNA looks intact and no RNA contamination appeared through gel image. Thus I don't know what kind of material within my DNA sample made Nanodrop overestimate the quantification.

I always used 5ug DNA as initial input for library. But now it looks I need a totally different way to quantify my DNA or I have some mistakes with DNA extraction?

Looking forward to your help, thanks
SK

[pmiguel]

Yeah, I have found it nearly impossible to remove all the RNA from a genomic DNA prep using traditional methods (like the ones you describe.) The RNAse degrades the RNA to short oligonuclotides that bind almost no EtBr (hence are invisible on agarose gels), but they still absorb UV during UV spec.

The EtOH ppt does fractionate away some of the oligo RNA, but by no means is it guaranteed to remove all of it. The issue is that you are starting out with so much more RNA than DNA in a typical cell (100x? 1000x?)

(But one major source of UV absorbance is phenol. Your precipitations likely got rid of it, but if even a tiny amount contaminates your sample, you will see very strong absorbance at 270 nm -- the shoulder from this phenol peak will confound your attempts to read DNA/RNA at 260 nm.)

So one method is just to run an agarose gel and quantitate based on it. Not extremely accurate, but it doesn't need to be. Alternatively, used double-strand specific fluorimetry.

We will frequently use Riboshredder on a DNA prep, then dilute the sample and remove the nucleotides using a microcon spin column. This will also remove any other small solutes that might interfere with later steps. However it will not remove large molecules -- like DNAses. So caution is warranted.

If you have a very large amount of genomic DNA, doing an RNAse treatment, followed by a precipitation can be very effective in removing RNA -- if the precipitated DNA is handled in a non-standard manner. Instead of spinning it down, the clump of DNA can be "hooked" (not spooled!) and transferred to a disposable test tube containing a few ml of 70% ethanol. It should be agitated a little then transferred to another test tube for another rinse. After this the clump can be transferred to a microfuge tube, spun, dried (minimally!) and then resuspended in your buffer of choice.

Obviously we are talking on the order of 100 ugs of DNA here.

Anyway, what you are running into is by no means an exotic problem. But there are lots of potential issues in play and I have found that the majority of protocols completely ignore these issues, or only concern themselves with one or two of the issues. Many robust protocols will work fairly well of a large range of starting DNA -- an order of magnitude at least, so you may have that going for you...

--
Phillip

[skblazer]

Many thanks again for your kindly response Phillip. You always answer our questions.

It seems I have to re-quantify all my DNA samples on a gel from now, or I have to purify DNA additionally using a column after precipitate, though it bring more loss.

For the Riboshredder you mentioned, can I just add it in tail buffer in digestion process or I have to use it for re-suspend DNA, I don't know if EDTA or Proteinase K will limit it.

Thanks,
SK

Yeah, I have found it nearly impossible to remove all the RNA from a genomic DNA prep using traditional methods (like the ones you describe.) The RNAse degrades the RNA to short oligonuclotides that bind almost no EtBr (hence are invisible on agarose gels), but they still absorb UV during UV spec.

The EtOH ppt does fractionate away some of the oligo RNA, but by no means is it guaranteed to remove all of it. The issue is that you are starting out with so much more RNA than DNA in a typical cell (100x? 1000x?)

(But one major source of UV absorbance is phenol. Your precipitations likely got rid of it, but if even a tiny amount contaminates your sample, you will see very strong absorbance at 270 nm -- the shoulder from this phenol peak will confound your attempts to read DNA/RNA at 260 nm.)

So one method is just to run an agarose gel and quantitate based on it. Not extremely accurate, but it doesn't need to be. Alternatively, used double-strand specific fluorimetry.

We will frequently use Riboshredder on a DNA prep, then dilute the sample and remove the nucleotides using a microcon spin column. This will also remove any other small solutes that might interfere with later steps. However it will not remove large molecules -- like DNAses. So caution is warranted.

If you have a very large amount of genomic DNA, doing an RNAse treatment, followed by a precipitation can be very effective in removing RNA -- if the precipitated DNA is handled in a non-standard manner. Instead of spinning it down, the clump of DNA can be "hooked" (not spooled!) and transferred to a disposable test tube containing a few ml of 70% ethanol. It should be agitated a little then transferred to another test tube for another rinse. After this the clump can be transferred to a microfuge tube, spun, dried (minimally!) and then resuspended in your buffer of choice.

Obviously we are talking on the order of 100 ugs of DNA here.

Anyway, what you are running into is by no means an exotic problem. But there are lots of potential issues in play and I have found that the majority of protocols completely ignore these issues, or only concern themselves with one or two of the issues. Many robust protocols will work fairly well of a large range of starting DNA -- an order of magnitude at least, so you may have that going for you...

--
Phillip

[pmiguel]

I don't know what your situation is. Are these fragment libraries, mate end or what? What method are you using to construct them? Hard to advise you without more information. If you have been successfully making libraries, then don't change anything. (You might want to run an agarose gel to make sure there is at least some genomic DNA there.)

Unless you have RNA adapters for some bizarre reason, I don't see any need to get rid of the riboshredder enzymes at all. They degrade RNA not DNA.

--
Phillip

[SarahNGS]

You may want to consider an alternate means of quantitation. NanoDrop quantitation, as stated above, suffers from contaminants. A dsDNA specific method like picogreen will tell you how much DNA you have in there regardless of RNA content. It requires a fluorescent-capable plate reader or fluorometer but it would be well worth it to gain access to one.

[niceday]

previously any samples sent to us that were quantitated on a nanodrop would be requantitated on site. We would never take a nanodrop value as accurate.

If you have always used a nanodrop without any problems then this isn't your problem.

We use a bioanalyser and a qubit.

You can use picogreen or quantifluor and a reader.

[rudi283]

I always use Qubit to quantify the input DNA, nanodrop only to check the quality

before the capture/sequencing I check the samples on qubit and bioanalyzer

[julihepple]

I also check the RNA content using a Qubit and after RNase 1 treatment, I use Ampure beads to clean up...seems to work beautifully every time.

[pmiguel]

Interesting. Ampure beads work fine for genomic DNA?

I mean I used to use a PEG precipitation method to "clean up" genomic DNA preps that did not want to digest with a restriction enzyme back in the day. So it stands to reason.

But any issues with there being too much DNA? Whereupon you could get a giant clump of "spoolable" DNA with your beads embedded in the middle? Or does that not really happen under the precipitation conditions Ampure uses?

--
Phillip

[cboustred]

Hi, apologies for tagging on the end of this thread, the current contributors seem to be very knowledgeable so I wanted to try and get their attention!

I have recently sent 24 samples (pools of genomic DNA, extracted via a salting out procedure that have been stored at -20, and freeze thawed multiple times) to our friendly University sequencing department for Illumina PE library prep and Nimblegen custom in solution capture -> Illumina GAIIx PE sequencing

I have been using these samples without a problem for standard PCRs in our lab.

I have standardised each of the samples at 10ng/ul (diluted in H2O) and measured the concentration on both a Nanodrop (Average conc = 9ng/ul, 260/280=1.9, 260/230=0.75) and a Qubit (Average conc = 8.2ng/ul).

The total amount of starting DNA per sample is ~1ug (the lower limit)

The first problem I notice is the average Nanodrop reading < Qubit reading, from my experience and what I have read, the Nanodrop generally overestimates DNA concentration relative to Qubit - any ideas why this is not the case here? Is the low 260/230 ratio significant?

The samples went through the first stages of library prep alongside other independently extracted samples.

The second problem was that the samples did not shear evenly or consistently using the Bioeruptor - taking > 40mins to shear to the required size - shearing checked on Bioanalyzer (normal and HS).

After an initial 8 cycles of LM-PCR the results (again checked via Bioanalyser) are very inconsistent, ~7/24 samples show a faint (expected due to low amounts of starting material?) smears of the correct size, the remaining samples are blank except for primer dimer / adapter.

A selection of the failed samples were re amplified for a further 7 cycles (15 in total) and run again on Bioanalyzer - the results show no improvement.

Any ideas what it could be that are bringing about these inconsistent results?

Any advice much appreciated

Chris

[pmiguel]

I think issues of this sort are usually caused by impurities in the genomic DNA inhibiting the first enzymatic reaction of library construction. If the fragments do not get blunted, the adapters will not ligate well.

Did you do any purification post-sonication?

--
Phillip

[cboustred]

Hi Phillip, thanks for your reply, I have passed it on to the Seq dept.

I believe the samples were purified down a Qiagen column post sonication. If the column isn't getting rid of the impurities any suggestions as to what would?!

Another suggestion I have had is that unligated adapter is being carried over into the PCR step and inhibiting it? Modification of the ligation and post ligation clean-up steps would be the answer if this were the case (I believe this step is bead purified)

Thanks again for your advice, very helpful

BW

Chris

[pmiguel]

Qiagen should be fine -- however any ethanol/guanidine isothiocyanate carried over from the column would be detrimental to downstream processing. So you want to make sure that does not happen.

I actually prefer Millipore microcon units to get rid of small molecule contaminants. You just dilute your sample to 400-500ul with water or EB, spin to 50 ul and repeat 1-2 times. Each time you dilute out small molecule contaminants about 10x. They do not remove larger molecules (many enzymes), but if you confident that is not an issue, they avoid the issue of adding contaminants back into your sample.

As far as unligated adapter inhibiting the PCR step -- that sound far-fetched to me unless you have an incredible excess of adapter that is not removed by Ampure.

One other factor that spring to mind -- are you controlling the temps in your sonication chamber? If not, you may be strand denaturing your DNA during sonication due to heat. Among other issues that might result from strand denaturing your DNA would be that T4 ligase would not work.

--
Phillip