Hi there,
I’m back for another question: Do you guys have any go-to techniques to replace gel electrophoresis? I usually extract up to 4k DNA samples in a good week or prepare as many PCRs as possible. I have to check each of them manually on a gel which consumes most of the time in the workflow at the moment. Are there good (and affordable) alternatives to that? I imagine the plate reader could be a good replacement for DNA extracts, although it does not give me any qualitative information. But how about PCR products? I need to confirm the right size of the amplicon… I know that the fragment analyzer is capable of doing this but it has insanely long runtimes and is expensive.
Only alternative that I’ve found that is fully automatable is the Fragment Analyzer. They do have a kit that brings the cost down to ~$50/plate. We use a 96-well gel that is ~$30/plate, so the cost difference isn’t huge. But the timing (45min vs 10min) is pretty big.
Everything else that I’ve found will tell you that there’s DNA there, but can’t confirm size nor “cleanliness” of the PCR product.
One option is the Ranger Nimbus. I don’t have direct experience with it myself, but they claim 384 samples in 30-90 minutes if you don’t need to recover the amplicons (it also has a size selection mode that drops the throughput to 96 in ~2 hours).
The Perkin Elmer LabChip GXII has been the gold standard for me for high throughput protein CE-SDS. Not fully end to end automatable, but it’ll chug through a 96-well plate in under an hour (about 1.5 min per sample). They are extremely finicky and chips fail all the time. But they are the fastest to result for any other instrument I’ve looked into. We’ve tried to integrate it with our Green Button Go workcell, but we couldn’t get access to an API and the best we could do is something like AutoHotKey to simulate screen presses to start a run. Literally no instrument feedback so if errors happen, you just got to be there to interrupt it…and bad runs happen all the time. You get about 4-8x 96-well runs before the chip dies. The consumables are also not robot friendly.
My experience with the DNA chip is a bit limited, but it is even more finicky than the Protein Chips. Nonetheless, the instrument is still the fastest/highest throughput out of all the instruments I’ve seen on the market and is probably still the gold standard if you want to move fast. Maybe some day someone will have a fully automatable instrument as high-throughput as the LabChip.
Other things to look into…Sciex PA800 Plus which also isn’t fully automatable, but is definitely the gold standard for later stage stuff. And Protein Simple iCE/Maurice…but I don’t think they can do DNA and are gears much more for proteins.
I only explored what I have available right now:
We have a Fragment Analyzer, although only with a 12 capillary array, which means one plate takes ~ 10 hours to measure.
We have a SpectraMax ID3 plate reader, this might be a great option to measure DNA extracts, maybe one can even come up with a custom DIY kit to reduce costs. It does not tell me much about quality though, on the other hand, I might opt for checking quality only for the samples that fail to amplify.
I have 4 96-well gel chambers, so checking 192 samples takes about 45 minutes to process and document. Material costs are close to 0 here, but working time in Germany is ~ 70€/h so I probably pay ~ 27 cents per sample.
So I’d need a solution in that price range but am already starting to realize that this might be impossible.
It does not need to be fully automatable. Our workflow at the moment prepares one plate that only goes to the gel per extracted plate, if I’d have a similar solution, I’m fine with taking out the plate and putting it in the next machine.
Looks like a possible solution, although 90 minutes for 96 samples is quite long. I’d need to buy at least 2 to not create a bottleneck. How about costs? I could not find anything on the net?
There’s a few different consumable for TapeStation depending on your needs. The most common that we use is:
$260/112 samples = $2.30/sample Plus other common consumables (tips, buffer, plates, etc).
So not super cheap, but it is walk away for a single plate. The data you get out is pretty similar to the Fragment Analyzer.
I think the step up that you could do is a FA with the 96 capillary set. This can do a full plate in ~45mins. But that’s a good bit of capital to purchase the instrument in the first place (~$80K IIRC).
Of course these are all US prices, no idea how that translates for the same consumables/instruments in Europe.
An additional cost element to consider for tapestation vs capillary electrophoresis is that it can run a smaller number of samples if needed while a frag analyzer always uses 96 capillaries.
Capillary arrays themselves can also get blocked etc whereas I think the tape station has less big budget maintenance items
Okay, I get the feeling that it is tapestation or nothing. I’ll take a look into this although I already dislike the costs and also don’t see any oportunity to do some costum stuff. Thank you all for your valuable input
Do you really need to check all of them? If cost is a concern, you could probably just take a subsample. If you’re running many different (or unknown) primers, this might not be advisable*.
If the primers are known and reliable it’s a very low risk proposition to just hit any pos/neg controls and 1-20 samples per plate and continue on if it looks fine. If you need to quantitate the DNA, there are plenty of fluorometric kits (e.g. Quant-IT) that can give you that information very cheaply.
*I’m in an early R&D environment so can be more loose with this sort of thing. If you’re in diagnostics…
For extractions: I could switch to a plate reader, and just measure all samples.
For PCRs: We do NGS amplicon sequencing after the PCR so I have to make sure that all PCRs worked. I’m not in diagnostics but run a core facility at a university and have all the freedom I want But since I pool ~ 1000 reactions and throw them on a NovaSeq which is not cheap I cannot afford to resequence samples out of laziness…
Ah yeah, I’m not encouraging laziness. Just considering the cost benefit since it was brought up. IMO it depends on how critical dropout prevention is. I’d argue with the larger sequencing pools that the cost per dropout is much lower. There’s obviously some delay to take into account between runs (e.g. you run the NovaSeq once per quarter) that could make a sample dropout more impactful to the folks looking for the data.
At some point, the cost of doing the QC is more than what it would cost to just resequence the dropouts you’d get. Go too far and your submitters may become unhappy and go somewhere else (or you lose your job I guess). Worth going through any historic data and seeing how many dropouts QC prevented.
Of course, losing some of your QC increases time spent troubleshooting and figuring out what went wrong so it’s not as cut and dry as saying we spent 5000 bucks to QC 1000 samples and prevented 4 dropouts which cost 70 bucks to in sequencing costs (or whatever it is).
Sorry for steering off topic. Back on topic, we use a tapestation for our QC. Moving to a fragment analyzer soon since it’ll be part of a larger integrated system.
We were in touch with the Tapestation PM, sales, and software team at Agilent regarding integration of the Tapestation. If you’re looking to somehow integrate the Tapestation / its software, be it for physical automation or just automated data acquisition, this is close to impossible and will remain so for at least the next couple years. Let me know if you want to hear more about that.
As a standalone device for automated gel electrophoresis, I’ve heard a lot of good things about it, but haven’t used it personally.
