I’m in a public molecular biology lab and we’re evaluating the purchase of a robotic liquid handler and I’d really appreciate feedback from people who actually use these platforms.
Our main applications would be:
• PCR setup (often with complex layouts)
• 96 and 384-well plates
• post-PCR work
• possible NGS library prep in the future, but not the main focus
Throughput is moderate: ~20–30k reactions/year.
One of our most important requirements is reliable pipetting at very small volumes (~1 µL).
Another key point is ease of use. The system will be used by multiple people in the lab, so we’d prefer something where protocols can be modified without needing a dedicated automation specialist.
Platforms we are currently considering:
Hamilton STARlet / NGS STAR
Tecan Fluent
Beckman Coulter Biomek i-series
Eppendorf epMotion 5075
Gilson PIPETMAX 278
For those who work with these systems:
How reliable is pipetting around 1 µL, especially in 384 plates?
How easy is it to modify protocols when layouts or reagents change?
Are there limitations that only appear after a few years of use?
How painful are the software environments in practice?
Regardless of the price, I would like to have some real-word feedback from anyone who used some of these platforms especially to check for any hidden drawback that would come up after some years of use.
I admit this isn’t a great solution due to cost, but for PCR setup for 384-well plates I’d be looking hard at an Echo for both primer and template dispensing. Can the other liquid handlers pipette 1uL? Yes, but if you need accuracy nothing is going to beat an Echo. I’m sure many others are doing PCR prep in the same volume range as what you want with a Tecan/Hamilton so I’d be curious to hear their experience with dialing in a 1uL dispense. I think once you have it dialed in it probably works pretty well. Getting it dialed in will probably take some time and pre-existing knowledge of the systems.
With that said, from your list I would only consider a Hamilton or a Tecan. And although I’ve preferred Tecans in the past their support (hardware and software) seems to have suffered since COVID. I’d probably be looking at a Hamilton now, I mean, just look at the support on this forum for an example as to why.
without needing a dedicated automation specialist.
No matter what you get I would highly recommend nominating someone whose job it is to be the guy for the instrument. If nobody is ultimately responsible it will become a mess and ultimately a really expensive (and heavy) paper weight.
Hi, the question is a little bit what your process looks like. Do you want to pipette 1 µL into a larger volume, or just the 1 µL? Single dispense, or multi dispense? If it’s single dispense into a larger volume, I think Hamilton, Beckman and Tecan can get you there. If it’s into a dry well or multi dispense, it gets tricky and expensive. You could go for Hamilton’s MagPip there, but I think your throughput is not high enough to make that worthwhile.
For software, general consensus is that Tecan and Beckman are more user-friendly for setup etc. while Hamilton takes some getting used to. On the other hand, Hamilton’s Venus will give you more freedom to do what you want (some on this forum might disagree, but I figure if you want ease of use you’re probably not thinking about employing Python et al.). Think along the lines of Windows vs. Linux.
As for the limitations, that strongly depends on where you are located. Some places have great coverage in service and application for their respective systems, others don’t. If you are in the US, Hamilton will have an extensive network of people that can help you. In Europe Hamilton and Tecan are a bit more on equal footing, for example.
Finally, I can also only echo that you need a dedicated person that takes care of the liquid handler. It’s an attractive idea to have everybody be able to do everything on the system, but the amount of lost knowledge when people leave (because everybody knows a little bit about everything) and the cost associated to mix ups (somebody opens up a file that the method of somebody else is currently using, causing an error), missed maintenance intervals (if everybody is responsible, nobody will schedule the visits) etc. far outweighs any benefits of that approach.
+1 to @jnecr‘s recommendation to consider an Echo if the cost isn’t prohibitive. Also agree that from the options you listed, it’s hard to go wrong with a Hamilton. Their hardware is top notch, though the programming does have more of a learning curve as others mentioned.
If you’re willing to look outside of that list but want cheaper options than an Echo, I’d consider an SPT Labtech Firefly. They come with a 384 head by default, and the positive displacement syringes might be helpful for your 1uL dispensing (if those dispenses are reagents and not samples). Their software is also pretty intuitive. Another positive displacement option is the Formulatrix Flo i8 PD, but I have no idea how it performs in the real world and it’s pretty new so I’d bet few people have experience with it.
Maybe this sounds biased coming from me, an automation specialist responding on a forum full of automation specialists, but you’ll need an in house automation specialist if you want the system and results to last.
Consider product development, process improvement, change control, and tolerance bands of any components in the assay from hardware to reagent lots - you’ll need someone who is able to manage and adapt not only the methods, but performance in response to uncontrollable variables.
At a minimum, I’d consider an ambitious early career employee with some engineering background and interest in automation, but if you highly value accuracy/precision and quicker or better documented development cycle, then you’ll want a dedicated person with experience.
For the other aspects of the question, I’d echo my colleagues. I personally have experience with Tecan, Beckman, and Hamilton and I’d consider 1uL to be near at the limit of volume ranges - especially with reagent like mastermix and possibly without a dedicated specialist. Liquid systems can be equipped with very low volume syringes (think of this like using a 1-10uL pipette instead of 1-1000) which will help but liquid systems are more susceptible to humidity and other environment factors and require more maintenance - daily, monthly, and yearly which sort of brings us back to having a dedicated person
Thank you very much to everyone who took the time to reply! I would like to clarify a few details.
We are a relatively small group (~10 people including technicians and researchers), and our throughput does not justify having a full-time automation specialist. Of course, we do have one person responsible for the instruments (maintenance, service contracts, troubleshooting, etc.), but the day-to-day operation is handled by all the technicians in the lab. This is the reason why I emphasized the need for an “easy-to-use” robot.
For context, we have been using a Tecan EVO for more than 10 years, so we are not completely new to automation. In practice we tend to use a relatively small number of protocols — let’s say around 10 — which would ideally be provided with the system. It is quite unlikely that we will need to make very complex modifications to them.
Regarding the 1 µL dispensing mentioned earlier: this refers to DNA samples. Typically, we dispense about 6–10 µL of master mix and then add 1–2 µL of DNA. I just wanted to clarify this point.
Is there something that prevents your current Evo from doing this work? What’s the Evo setup? Liquid displacement or AirLiHa? Multi-channel arm (MCA)? There’s nothing ground-breaking on the newer instruments that will make the 1uL dispense way better, depending on your current Evo’s setup.
Hi, I work for Gilson as an applications expert supporting customers. I’m glad to see that you’ve listed the Pipetmax 278 as one of your potential options. I wanted to check that you’re aware of the new application based packages that we now offer, including one specifically for PCR plate setup - qPCR - Application Center.
I also want to draw your attention to an application note we have here -AN1053_THREE_APPROACHES_TO_QPCR_REACTION_SETUP.pdf - which focuses on the comparison of the Gilson Pipetmax systems with manual pipetting (by a very well trained hand). You will see that the results are excellent either way and very comparable. This is in part due to the fact that the Gilson automated systems use the same technologies as the manual systems (same pipette head sizes (e.g. P20, P200 etc), same type of disposable tips, same air displacement tech (e.g. pistons and seals etc). So you can move from manual to automated with no deviation in results and you can also switch to and fro as well, as your workflow or workload requires.
As for ease of use, this is always going to be a personal thing. Certainly there is no requirement for scripting or python. Everything is drag n drop, but as always, everyone has their own preferences as to how the software functions. We would provide some example protocols with the system, including one that is custom written as per your request. And the package includes wizards specifically for creating qPCR protocols and also for doing normalization. These are very easy to use.
For pipetting specs you can look here - gb.gilson.com/GBSV/amfile/file/download/file/2218/ - as others have already mentioned, you will pipette the 1 uL better if it is into liquid. So for PCR you will want to pipette the MM first and then the DNA sample and mix without changing the tip. Try this by hand with a pipette (assuming it’s good quality and within calibration) and you should be able to achieve the same with a Pipetmax.
Hope that helps, from the Gilson Pipetmax 278 side of the discussion. I’m UK based but if you want me to put you in touch with my Italian colleagues then I can do so.
There are questions you should consider outside of your original post. The i7, Hamilton, and Tecan on your list will all do the liquid handling about equally well and there is more to evaluate than the robotics. You’re thinking long term, so you might consider your strategy to digitally integrate these systems to handle sample intake and tracking. CSV file worklisting will be the best solution unless your lab decides to buy an enterprise LIMS system or develop a lot of python scripting.
The software environments are painful for all of these systems. You don’t need a full time specialist, but you will need someone with the time and skills to create and manage methods, calibrate liquid classes, and do some scripting depending on overall workflow.
How reliable is pipetting around 1 µL, especially in 384 plates?
The scenario you describe (1ul transfer, mix into master mix), is on the edge of reliable with careful calibration. You’ll get <10% CVs if done well. Going to 2ul will get you a massive improvement.
How easy is it to modify protocols when layouts or reagents change?
If you expect regular change, protocols should be set up as templates, and user inputs at run start should be changing liquid classes, layouts, etc. Hamilton Venus will be better for this.
Are there limitations that only appear after a few years of use?
This depends as there are so many ways to expand the automation robotics, integration, and software.
How painful are the software environments in practice?
All painful until you get used to it. These are all running on old software updated over decades. If you can demo the Firefly, you’ll see a much more modern software that’s truly easy to use. But it doesn’t have 8-channels and might not work for your DNA transfer.
Your lab should consider whether to get a “traditional” liquid handler on your list at all, or get 2-3 small instruments. The Mosquito would handle the 1ul sample transfer very well. You can get one of many reagent dispensers of 96-head stamp instruments to do the rest. This can be a cheaper, more versatile, and much more user friendly automation setup. I’d recommend this path based off the context given.