I'm having troubles with some DNA extractions. The yields are much lower than expected, and the Nanodrop spectrophotometer is showing quite a bit of contamination through the 260/230 ratios (see table below). The samples consist of mixed marine fine-mortarted fauna (environmental samples), of which I want to sequence a couple of marker genes to determine species composition (metabarcoding), using paired-end Illumina MiSeq. I could just go on with the PCR as I'm supposed to and see if that works, but my concerns are still:

  1. As the DNA extracts contain an order of magnitude less DNA than expected, would you really consider them representations of their whole communities?
  2. Assuming the PCRs work out and I get nice bands, is there still a risk that the contaminants could disturb the sequencing process? That is, given that I only input 2 out of 25 µl of DNA extract into the first PCR reaction, and just as little PCR product into the second (index/library prep) PCR. Can the Illumina flow cells be sensitive to these contaminants as well?
Sample ID Nucleic Acid Conc. Unit A260 A280 260/280 260/230 Sample Type Factor
1 7.6 ng/µl 0.152 0.086 1.76 0.71 DNA 50
2 24 ng/µl 0.479 0.249 1.93 1.16 DNA 50
3 21.9 ng/µl 0.439 0.238 1.84 1.15 DNA 50
4 18.6 ng/µl 0.371 0.190 1.95 1.29 DNA 50
5 18.5 ng/µl 0.371 0.200 1.85 1.44 DNA 50
6 6.9 ng/µl 0.138 0.069 2.00 0.70 DNA 50
7 7.9 ng/µl 0.157 0.092 1.72 0.78 DNA 50
8 35.2 ng/µl 0.703 0.363 1.94 1.67 DNA 50
9 9.6 ng/µl 0.192 0.099 1.94 0.85 DNA 50
extraction_blank 1.6 ng/µl 0.031 0.006 4.84 0.19 DNA 50
collection_blank 3.9 ng/µl 0.078 0.036 2.19 0.42 DNA 50

According to ThermoFisher "good" values are between 1.8-2.0 for the 260/280 ratio, and 2.0-2.2 for the 260/230 ratio. Less may indicate the presence of salts or EDTA for example. The concentrations are for 100 µl of DNA extract, that is, total DNA yield is given by multiplying the concentrations with 100. For the record, I have heard from people who get 200-300 ng/µl DNA concentrations from marine sediment extractions. My samples should in theory contain even more than that. Although I have earlier only gotten concentrations in the range of 30-100 ng/µl for the same type of samples with the same extraction kit.

What would you do in this situation? How do you go on with values like this from the spectrophotometer?


1 Answer 1


I have never performed environmental extractions, and I am not familiar with your specific protocol. I would defer to any other commenters with more experience with these methods. However, I would recommend redoing the extractions, if it is feasible. Any time a standard protocol gives results that are significantly unexpected, it's worth thinking hard about. In this case, where it is your initial step, it's even easier to suggest, since there's much less wasted time than if it were later in your protocol. The quality of your data depends on what goes in, so it's not worth risk it.

Nanodrop quantification is not particularly reliable for low concentrations like this. That makes the ratios even more questionable. I would not try to make too much of these particular metrics beyond "this was a poor extraction."

PCR could certainly give you plenty of DNA even from a minuscule sample - that's the whole point of exponential amplification! But if your extractions are poor enough that they are extremely biased, that would carry on through PCR.

I would worry less about contaminating the sequencing reaction at the end, but it's still worth trying to get pure DNA. What's your cleanup before dilution? Anyway, if you do have an inhibitor, I would expect it to give issues during PCR. And you'll be diluting them 20? 100?-fold for loading onto the flow cell.

  • $\begingroup$ Hi. I forgot to add that I've redone the extraction four times already, following the kit manufacturer's troubleshooting guide for low DNA yields: heating up the elution buffer, letting it incubate for longer, running the elute through the spin column multiple times. I've even bought a new kit to try with. The values above are as good as it gets. Thought I was gonna do another thread about that... And I've double-checked the concentrations with Qubit, although not for this last specific extraction attempt. It doesn't differ that much. $\endgroup$
    – Joel
    Commented Jan 9 at 7:01
  • $\begingroup$ The point of this thread was more to check which possible unforeseen problems I might encounter later, should I choose to continue with these DNA extracts. I perform post-PCR cleanup using magnetic AMPure XP beads. The final dilution of the DNA extract in the library will probably be closer to 1:100. But I take your answer then that I should not worry about contaminants as long as the PCR bands look nice – I should worry about bias from the low DNA yields? $\endgroup$
    – Joel
    Commented Jan 9 at 7:10
  • $\begingroup$ With a bead clean and that dilution, I wouldn't worry about sequencing. I would still be a bit concerned about bias. I don't know how quantitative you need this assay to be (do you want to measure frequency or just presence?), but I would be concerned by the extraction issues. $\endgroup$
    – ksdjnf
    Commented Jan 9 at 14:58
  • $\begingroup$ In metabarcoding studies, quantitative (read count) measurements are largely useless due to the large bias introduced by PCR. Presence/absence is the primary goal. I made another thread about the low yield issue: biology.stackexchange.com/questions/113837/… – Kudos to whoever can solve that mystery. $\endgroup$
    – Joel
    Commented Jan 9 at 15:29
  • $\begingroup$ You might be able to get away with using these, then, but I would definitely qubit instead of nanodrop and be skeptical about any results you get. $\endgroup$
    – ksdjnf
    Commented Jan 9 at 15:39

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .