So, my friendly Doper molecular biologists, I’ve mailed in an aliquot of DNA to be sequenced by a commerical sequencer but unfortunately it seems to have failed. Base upon the chromatogram is there anything that you Dopers might guess have caused it to fail?
I know that images of the chromatogram can occassionally be useful in diagnosing the problem: too much DNA, too little DNA, etc., so if you’ve got the molecular biology/biochemistry skillz, take a look and I’d be much obliged:
Disclaimer: I’m still a student. I’m graduating in a few weeks, but still a student. A bored student who is procrastinating. Anyway, I’m taking this as an academic exercise for me, so take me with a grain of salt.
Your last Chromatogram looks halfway decent compared to the rest of them. Was that sample in anyway different from the others?
Things I can think of…
was the package you mailed out iced?
contamination
If you used the wrong primers during the PCR process then you’d get gobbly-gook at the end.
Primer annealing temperature was too high.
Extension phase was too short
All those other PCR temp/time variables
If the sequences are very very short(short enough for spontaneous polymerization of nucleotides) or very very long(long enough that gel electrophoresis stops being linear) the sequencer won’t be able to handle them.
This here is complete speculation on my part. Too little DNA would only be a problem if it’s below the threshhold for reliable PCR. The primers can’t anneal to the DNA quickly enough and you can have no or only partial extension during replication. When this happens you will get many different sized fragments, but most of them will be toward the lower end in size. What this means is that the end portions on both ends of the DNA sequence will look like gobbly gook, but the middle portion might look OK. (note: there is a goldilocks zone between a bunch of variables including the kinetics of primer initiated extension and fragment initiated extension. Fragment initiated extension can cause the whole graph to look like gobbly-gook).
Again, speculation. In the case of too much DNA there would be no lack of template DNA for the primers to anneal to. Primers would anneal immediately and if extension phase is sufficiently long there will be complete replication. However, with so many replication events happening at the same time there might be competition for nucleotides. If this happens you’ll end up with many aborted fragments of completely random sizes. Your entire graph will look like gobbly gook.
The easy way to test this is just to send in serial dilutions.
uh I just realized, if the DNA sequencing machines uses Celera’s method of shotgun PCR and using a computer to sort the overlaps… then 8 and 9 are based on the wrong model and are completely irrelevant.
Heh, if I could only spend this much time on my schoolwork…
Well, all of those things Harmonix said could cause the sequencing to fail. The chromatogram generally tells you very little, and your chromatograms look to me like the hundreds of failed reactions that I’ve seen. The new ABI machines run with the new BigDye are almost completely automatic. If you are sequencing with standard primers (T3, T7, Sp6, M13) then there is even less chance of failure. So generally speaking, there is no PCR variability involved.
The usual reasons for sequencing to fail, in my book, are
poor sequencing primer magic (if not using standard sequencing primers) – solution is to design new ones. I like using 21-24-mers, GC content 45-55%, no inverted repeats > 3 bp, ending with 2 As or Ts. These are the genome center rules and in my book they seem to work better than 3/4 of the time (for PCR and for sequencing).
The DNA requirement is usually pretty flexible, just so long as you get around 50-500 ng of template in the mix, you shouldn’t have a problem.
Poor cleanup of the post-sequencing sample.
A little more information is needed to fully understand what is going on here: are the three chromatograms using the same primer, the same template, are they redundant reactions? Have you confirmed the template is correct using basic restriction mapping? Are you providing primer? If so, has anyone recently sequenced successfully in the lab using the primer?
If you are running the reactions and having them reading them, it introduces a whole host of new problems. The BigDye can be bad, your BigDye may be different than what their machine is set up for, they may have their machines rejiggered to use different amounts of BigDye (we used to sequence using 1/2 the amount of BigDye recommended, 1 ul in 10 ul of reaction – the genome center at our school sequenced with 1/8 BigDye recommended). The reaction may have been run using the wrong setting (it is not a standard PCR setting – short denature, short annealing, long extension). And finally, you may have screwed up the cleanup to remove unincorporated dye. They say you can do it with a quick isopropanol precipitation but in my experience it is far better to use Sephadex G50 columns.
The easiest thing to suggest is to reprepare your template (I assume it was a miniprep or something easy to make), confirm with a few restriction cuts (pop out the insert, two or three directional cuts), and resend. If you are using custom, unproven primer, you may want to redesign.
Okay, sorry everyone, I should have provided a bit more information.
First, I think the basic answer is that unfortunately the chromatogram can’t tell you much about why a set of sequences failed.
Also, we submitted a control of a plasmid without our insert target that basically came back okay.
Right now our most likely cause of failure for this sequencing was due to choosing a primer binding site that was too close to the place where we inserted out vector and we think that our primer binding site got screwed up. We might try this again with another primer site that is further upstream and less likely to be disrupted by our insert.
Contamination is another possiblity that we’re trying to chase down but we don’t know where it might have come from and again our negative control seemed to sequence fine so that contamination must have been limited to only the three samples that I posted.
So, thank you guys and I guess we’ll be chasing down Harmonix’s 2 and 3 and edwino’s 1 and 2.
Harmonix, we actually mailed these suckers to your hometown, Davis Sequencing.
edwino, you’re an MD or MD student, correct? Are you in an MD/PhD program or just that smart?
Yup, defended my PhD thesis in April 2005, got my MD last Tuesday. Onto being a doctor for a while. When I get back into research in 3 or 4 years, I doubt there will be a lot of capillary sequencing still being done.
During my PhD, I sequenced around 1 MBp of my own data – mainly through shotgun sequencing some BACs and fosmids that I cloned, all the rest through old skool sequence, design primer, sequence, lather rinse repeat. In the begining, I poured my own acrylamide, ran my own individual [sup]35[/sup]S labeled dideoxy termination reactions and read it in using a tablet and a light pen.
Okay, pops, and did ya’ have to walk uphill both ways to the lab in three feet of snow?
I’m kidding, of course, but congratulations on both the MD and the PhD! I think that’s an almost unimaginable amount of schooling and work but congratulations. So I take it you intend to return to research work after your residency. If you don’t mind my asking, what were/are your research topics of interest?
I was pouring acrylamide and doing my own sequencing until, IIRC, 1997.
It was an unimaginable amount of school. When I graduated on Tuesday, it concluded basically 27 years straight of school, from nursery school in Skokie to primary education in Houston to 4 years of college in Austin back to Houston for 9 years of post-bac.
My PhD work started out in pretty straightforward developmental genetics, with the genetic regulatory networks involved in the development of the fruitfly (Drosophila) compound eye. Around the middle of my PhD, I began to become interested in the regulatory elements of these genes, namely the upstream enhancer elements. We decided first to try and use conservation between different species of Drosophila as well as binding site prediction to predict upstream enhancers. We had mixed success. This kind of evolved into my thesis work, which was (finally) published in Genome Research a few months ago. We did a pretty cool project where we predicted downstream targets of the gene eyeless first by constructing and predicting Ey binding sites across the Drosophila genome, then by using microarrays of different genetic backgrounds (over- and ectopic expression, as well as mutants) to find eye genes responsive to eyeless expression. The overlap of these was a compact list of 22 genes. Out of these, I showed that some of our predicted Ey binding sites were indeed functioning both in electromobility shift assays and chromatin IP as Ey binding sites. Better yet, DNA around the binding site drove eye expression in reporter constructs and responded to ectopic eyeless expression. So a microarray/biochem/genetics/bioinformatics type project.
In the future, I’m getting out of the fly business but I plan to take this approach to other systems. My residency will be in Internal Medicine. I’ll do a research-based fellowship, perhaps in Infectious Disease or Endocrinology or Pulmonary/Critical-Care. I’m still thinking about exact research projects to do, but the paradigm I am operating with is that disease can be treated like a developmental process, and dissected as such. Now we have tools for doing this across the whole genome from clinical samples. Specifically, I believe that we can examine the genetic regulatory fingerprint of diseased tissue samples using microarrays and stuff like ChIP-on-chip with neighboring less or unaffected tissue as a control. We can also do this across the process of disease development, so for instance with the first signs of insulin resistance until the development of type II diabetes mellitus. Or pneumonia into ARDS or sepsis. I believe that the regulatory profiles of these will be useful as a diagnostic and prognostic tool in itself, as well as giving us insight into the regulatory pathways and the networks changing in the disease process. These in turn may also be useful diagnostically, prognostically, and eventually, therapeutically.
Whew. I tend to get overexcited about this stuff. Guess I’m still a true believer in the physician-scientist thing, even after 9 years of school. Let’s just see if it remains pragmatic when I get out there in the real world.
Well, that sounds like some pretty interesting topics. I’ve kind of hoped that my lab might be more interested in bioinformatics and some other more high-throughput techniques but it’s a very structure oriented lab and right now the project that I’m working with my post-doc is pretty much bah-dum-bah trying to get soluble protein.
I hope that you can apply this to disease-processes like you’re hoping and it seems to have a lot of potential but those clinical applications are always so far away, aren’t they?
My goal is just to learn something useful. I don’t have that much interest in the licensing and development of products. But I think we are pretty close to being able to do genetic experiments with near-direct clinical relevance so perhaps the turnaround can shorten.
Anyway feel free to email if you have any questions or suggestions!
I was using the same primers to sequence as I had used to clone the gene in the first place. The primers were, apparently, not the most specific, and had amplified several different sequences. When I cut the DNA (which I thought was a single gene, but in retrospect was not a clean band) from the gel, and cloned it, I had cloned in several different sequences.
Thus, when I went to sequence (using the same primers), I ended up sequencing three or four different sequences simultaneously, resulting in the crappy chromatograms.
Move your sequencing primers in or out, and see what you get.
I jumped to industry as soon as I got my PhD two years ago.
It’s actually not as bad as they make it sound in grad school. I rather enjoy talking to the lawyers and the development people. Learning something useful and development don’t have to be mutually exclusive.
This is sort of a crusade of mine to open the eyes of graduate students to looking at positions outside of academia. More than 90% of them are not going to get that tenure track position, so I wish that the academics would stop bad mouthing anyone who leaves academia. Academic science has become something of a Ponzi scheme, and people would be foolish not to keep their options open.
FWIW, I’ve sequenced genomic DNA off of PCR products using the same genomic PCR products with minimal cleanup (just ran it through a Millipore column to get rid of primers and then used a few microliters in my sequencing reaction). The reads aren’t as long, but you can get 500 bp off of them. You just have to have damn good primers – granted I was using a low-complexity Drosophila genome template, but I still did BLAST searches against with my primer sequences and discarded anything with greater than 14 matches in a row at the 3’ end. One has to also be very careful about PCR, using protocols like touchdown to make sure you have good specificity.
Fiveyearlurker – I’m not one of those PhDs with anything against industry. My wife, an OD/PhD, is seriously considering industry jobs currently. I’d definitely consider it if I could do what I wanted to do – I’m just most interested right now in the hardcore bioinformatics/genetics aspect of disease development. I’m sure I could do that in industry as well, but my interests are less in the product development/FDA approval/patent law aspect of things but the science. That’s not a knock against industry at all, I fully recognize the necessity of all of those things, and they may become very appealing in the future for me. Especially given the huge overlap of clincally relevant science in industry and academia. I think the default route especially as an MD/PhD lies in academia at least until sometime during my junior faculty appointment (let’s say the next 10 years), but after that, I’ll go where I can do the best work, even if that means industry, leaving the country, or even (gasp) leaving science altogether.