The technicians reported ‘poor resolution’ for my samples. Why?

Loss of Resolution

If the Sequencing Techs say of your sample something like: “poor resolution from start” or “loss of resolution after 220 nt”, you need to read the following description of the problem and suggestions on how to prevent it.

DNA sequencers based on capillary electrophoresis are sensitive to the presence of certain contaminants that were not a problem with the old gel-based machines. These contaminants may not affect your sample in any way other than sequencing – you may be able to PCR them, restrict them or even transfect them with no problems.

We do not know the exact nature of these contaminants, but we know how they typically get into your samples, and we know how to remove them.

Here’s a very typical example of a Loss of Resolution (LOR) sample:

Early in the run, we see reasonable resolution:

50 or 60 nt further on, it’s obvious something is wrong:

Another 120 nt, and the resolution is terrible – the sequence is essentially uninterpretable:

We have strong reason to believe this is caused by an unknown contaminant in your template DNA.

Read the section at the bottom of this page for why we know this to be true.

Here’s what you can do to fix the problem:

  • Clean up your bad samples:

    Here are three suggestions for removing the unknown contaminant:

    1. Perform a phenol-chloroform extraction on your template DNA.
    2. Perform ammonium acetate-isopropyl precipitation on your template DNA.
    3. Pass your sample through a Sephadex G50 spin column.


  • Don’t let the contaminant get into your samples in the first place:

    Mini-preps are the most common LOR samples, and many Cores have observed that overloaded Qiagen preps (or similar silica preps) are at fault. When you overload these plasmid prep kits, they leave many impurities in the DNA. Do at least one of the following:

    1. Grow your minipreps for a shorter time. 12 hours is good, 8 hours is better. Yes, your yield may be reduced. The DNA will sequence better, though.
    2. Grow smaller bacterial cultures. Be conservative! If the kits says no more than 5 ml of lysate, limit yourself to 3 ml! Yes, your yield will be reduced. It’s worth it.
    3. I’ve heard rumors that certain E. coli strains are particularly problematic – specifically the ‘JM’ strains (JM101, JM105, etc). Be especially conservative when prepping plasmids from these strains.
  • The Sequencing Core is working to help on our end as well:

    I mentioned above that Sephadex G-50 spin columns can remove the mystery contaminant. We in fact use Sephadex as part of our sequencing protocols. Ours are optimized primarily to remove unincorporated dye-labeled nucleotides, but we have tested some protocol changes and alternative column sources that seem to minimize the incidence of LOR. Improvements since February 2003 have primarily been due to this effort.

    DO NOT count on us solving this problem! It is primarily a template purity issue, and we cannot promise to get good sequence from bad templates.

For those who want more information on this artifact, what causes it and what prevents it, you may want to continue reading.

There are four known causes for the LOR problem.

Three of the failure modes are within the Core itself, and are relatively uncommon. We watch for them all the time. The last failure mode is very common, and is a characteristic of entire sets of samples. In other words, it is a sample prep problem on the part of our client.

The known causes of the LOR problem are:

Air bubbles in the electrophoresis capillaries
If a capillary is partially blocked by an air bubble (which occasionally occurs during filling of the cap with electrophoresis medium), the sample in that capillary will exhibit extremely poor resolution (if it elutes at all).

This results in sporadic failures of individual samples out of sets that otherwise gave good results. The Core will always repeat samples exhibiting sporadic LOR.

Plugged cuvette flow ports
The flow paths in the instrument can become blocked, in which case an entire set of samples will fail with loss-of-resolution – every sample, including standards.

We have not seen this problem in our lab.

Overloaded lanes
We have been told by other Core facilities and by the sequencer manufacturer that excessive amounts of sample (sequencing termination products) in a capillary can cause a loss of resolution much as we describe.We have never seen this effect, and in fact have tried to reproduce it with no success.
Contaminated samples
We have time and time again seen entire sets of samples produce LOR, when the rest of the samples in that set (including standards) look fine. Samples such as these, when re-run, always give the same result – LOR. When the samples are cleaned up (e.g. by phenol-chloroform extraction or ammonium acetate – isopropyl precipitation) they produce excellent results.

We can only conclude that there is a contaminant in the samples that reduces resolution in our machines.