Anyone have any experience with automated colony pickers? Have some tangential experience with QPix but curious what other options are available on the market.
I found this iGEM project where an OT-2 was retro-fitted to be a colony picker and have toyed around with the idea of doing something similar on a STAR:
@LukeWitt - Hamilton offers EasyPick, currently a STARline product, that enables automated colony picking right on deck of a STAR, using channels and tips for picking. I believe this will also be available for VANTAGE soon. It requires addition of an additional channel to the pipettor (outfitted with an integrated USB camera), as well as a controllable light table carrier on deck to provide backlight and contrast for imaging. EasyPick includes a VENUS driver, as well as additional software and machine vision for runtime analysis and processing of images.
Several petri dish styles are supported, and full configuration of colony analysis and picking criteria is enabled through the EasyPick software used by VENUS.
I have used EasyPick reliably for various picking applications over the years both in my customer days, and as part of the Hamilton applications team. Applications ranged from simple for simple to complex picking protocols. Depending on the type of labware used, very high throughput and substantial walkaway time can be achieved.
I always appreciated this being fully integrated into VENUS, as it gives full control of customizing features of the application per the protocol requirements right in the method/workflow, not to mention being able to apply barcode/sample tracking metadata to the picked colonies and their labware destinations, for robust traceability.
Here is an application note with some more information. Let me know if you have any questions!
Singer Instruments has really, really good documentation for their cell picker API. Never used them but the interface seems great.
API docs here:
http://si-pixl-client-documentation.azurewebsites.net/webframe.html#topic1.html
SciRobiotics partners with Tecan for a very elegant, automated colony picker.
Check out the link below:
I have used the singer instrument. Its a nice idea to use extruded 3d printer plastic to pick colonies. On most platforms ive used picking is not the problem, its getting consistent image segmentation. Always seems to require tweaks but prob something you can get right if using regularly.
The Tecan Omni colony picking add on came to mind, like what Joan said. Automation removes the colony-picking bottleneck - The Blog - Tecan
Reviving this to see if anyone can help.
What type of picking system did you eventually go with @LukeWitt ?
Has anyone had issues with picking getting “lost” in the middle of the run. Our Pickolo frequently errors with a fatal XY while picking.
We went with a Qpix! It’s an overpriced one-trick pony and it’s incredibly locked down. You have to involve Molecular Devices for any and all modifications to the built-in routines, and they’ll charge you 9k for the pleasure of waiting several weeks to months for the development.
The first QPix we bought directly from Molecular Devices had a bad board that caused height sensing and barcode scanning issues that took them over 6 months to diagnose and replace.
Since the new instrument was such a lemon, we figured we’d save ourselves the hassle and purchase a second one used at auction with all the software and hardware required to run it. Once we got the used instrument, Molecular Devices told us that because the software EULA is non-transferrable, we wouldn’t be able to buy parts or get support without paying them an extra 14k for a brand-new license, and 9k per “extra” picking routine (most of which were already on the computer when we bought it), plus 4k for a new computer. We finally caved and paid, just to have them delay shipping the new computer for nearly a month. We submitted the PO back in mid-March and as of writing this, we’re still waiting for the computer to ship.
That’s just my experience, so take it as you will, but you’d have to pay me to buy another Molecular Devices instrument of any kind.
Hi.
Thank you for the feedback on the QPix. I am sorry that it’s been an expensive lemon. It’s another company to be added to the “no fly” list.
Depending on your needs you might be interested in Singer’s Pixl, the picking itself is quite a bit slower than for the Qpix but I know several people with other Singer devices and they’re very happy with them.
Thanks @JBagley, we have the Scirobotics Pickolo incorporated with our Fluet. It has been a bit buggy, but we’ve maged to sort most of them out.
Well we currently have a Qpix that hasn’t worked reliably since October of last year and we’ve already had them come by twice only to encounter another problem (we don’t have a service contract). But it’s also 10 years old, so we have been seriously considering getting a new one 
. It’s not easy finding a system with the same kind of speed. What version(s) did you get @LukeWitt ? PIXL is ok, just not very high throughput. Has anybody ever seen or used a K8 colony picker?? Made by K biosystems…
Wow, this sounds exactly like any other Danaher company. Danaher takes great companies with great products and picks them to the bone leaving nothing. I fear Beckman is headed in the exact same direction.
I’m curious how you are finding the Pickolo integration to be on the Fluent?
What tips do you use for picking? 10uL DiTis? Do you have problems with alignment (outside of the XY errors you mentioned)? And this is a very naive question, does the tip just touch the colony? Or is there an actual aspirate/dispense step involved?
We’re investigating doing automated picking of colonies into a 96-well plate and also into a PCR reaction to generate linear DNA. On the surface it seems like a Pickolo on a Tecan would be great for this.
We’ve had a few glitched such as communication errors that needed us to add dedicated USB cards to the PC as well as having to tweak the spatial parameters, but we are happy with it. The commands are part of the Fluent Control interface and are easily included in a script with loops. We have a robotic arm and space for 60 plates in stacks meaning we can have 30 plates with colonies brought to the Picolo and then transferred to “used”.
Regarding tips, we are using the 200uL DiTis and we have had a few missed picks. Changing to the 10’s might solve this but they don’t fit into the 2.2ml deep well plates we use for culturing. There is an actual aspirate /dispense process and we mitigated the accuracy issue by making the tip do a small movement (x,y,-x,-y) once it has contacted the media. The dispense command also includes a “mix” step.
automated picking of colonies into a 96-well plate and also into a PCR reaction to generate linear DNA - this is exactly what we are using the platform for, along with a lot of other functions. Scirobotics also has a automation option for spiral inoculation.
Please give me a shout to compare notes.
Many thanks for your reply, this answers a lot of my questions. Sounds like the Pickolo is a pretty customizable solution to mitigate common errors. My biggest concern is missed picks due to how DiTis are mounted on the cones. They frequently get a bit of a snaggle-tooth. I’m hoping we can use 10uL DiTis which should mitigate a lot of that issue, but even if we can’t we can probably get by with other solutions to that problem.
Has variability in agar thickness been a problem? Do you pour your plates in house or order pre-made plates?
@jnecr The pick and move function that we created a subscript for solved the missed picks for us. It is very customizable.
@Alasdair we correct the Z-position for the agar thickness. We are currently pouring our own plates. We had to make a granite surface in our laminar flow cabin. The plates have to be very flat and fairly consistent height. There is also a Z-position offset in the picking script to make small adjustments before the run starts.
I assumed you could use liquid level detection to just detect when the tip hits agar, is this not the case?
