Robot for pooling/normalisation

1

Hi all,
I was wondering if you can give me some opinions on a good robot for pooling/ normalisation. Ideally I would like a robot with separate channels that can pipette different volumes for pooling (plate into tube) and normalisation (plate to plate). Ideally the robot also has a 96 pipetting head so I can employ it for other processes too. I know the Hamilton Star is an option. However it is a quite big robot…is there anything smaller and cheaper that you can recommend?
Thank you for all your help!!

2

What sort of budget / size / other constraints do you have?

Hamilton, Tecan, Beckman Coulter all have offerings here. Opentrons flex is likely an option for you to check out but I haven’t ever used one in the lab.

Smaller deck size can be fairly limiting (e.g. a dual armed starlet) but is workable if you are flexible with your walk-away ability or processing plate number. Especially if filter tips are needed.

3

There are also 96 heads that allow you to pipette varying volumes and so effectively you can in theory use them like an 8 channel.

4

Hello Luis,
thank you for the post. Do you refer to the Lynx by dynamic devices? Have you tried it?

5

Hello mnewsom. Thank you for your help! I will check out the robots you mentioned.

1 Like
6

I think they are the only independent volume 96 head on the market. Their systems have some other neat advantages too (dynamic swapping of heads between 24/96 during the run, positive pressure potential through the head), depending on what else you’re looking to do.

OpenTrons will most likely be your cheapest/smallest.

1 Like
7

Check out the Agilent bravo too. it has a 96 Channel head that can pick up 1 tip at a time. This is slow for normalizing a 96 well plate (45 minutes) compared to a multispan. You will also get more for you money if you purchase an NGS option A as it includes a shaker, magnet and 3 temp controlled positions in a very small footprint

2 Likes
8

Hi Marcella,

I would also start by asking what your budget is, what you estimate your throughput to look like, do you know (some) programming, how much automation experience you have and how much time you are willing to invest into your automation future.

One consideration: normalisations and pooling normally do not require very fast processing but can potentially require very complex pipetting patterns (depending on application and throughput).
Though in my experience, 95% of the time, all you want the liquid handler to do is to take input= 1)list of source positions, 2) list of destination positions, 3) list of volumes to add (probably divided into a. diluent and b. sample).

I have used the Opentrons OT-2 for years to do exactly that.
The fact that the OT-2 only has a single channel/pipette was actually not a handicap but a feature because it meant the programming was absurdly easy.
When adding the diluent first and you use “aliquoting”, “multi-dispense” or “distributing” (in Opentrons lingo) this is also incredibly fast, and the Opentrons API has a smart way of self-refilling the tip if the total volume of transfers exceeds the max tip volume which works very well.
Furthermore, the OT-2 might be one of the most precise out of the box liquid handlers there is. For almost all other robots you have to tweak your liquid classes (i.e. the “bag of pipetting parameters” as a friend of mine likes to call it) quite a bit.

That being said I have made our Hamilton STARlet (the STAR’s little brother) work out fast aqueous-based, 8-channel normalisations with high precision and accuracy, and you might want to have a look at that system with the 96-head.

But the only way I was able to make the STARlet perform liquid transfers like…

source_wells = [
    plate_1["A1"],
    plate_2["C6"],
    plate_2["H3"],
    plate_2["B10"],
    plate_3["E3"]
]
transfer_vols = [
    120,
    85,
    60,
    190,
    140
]
dest_wells =[
    plate_5["A1"],
    plate_5["C1"]
    plate_6["E5"]
    plate_6["F7"]
    plate_6["A12"]
]

…i.e. from any source position list, any liquid volumes desired to any destination position, is with PyLabRobot.

I should mention that the above is a heavily simplified example.
When you have independent 8-channels multi-dispensing with each channel dispensing a different amount in every well - something very likely to occur during normalisations - it is actually quite a complex task to programme this without a programming language.

If you are willing to learn (or already know) some Python then PyLabRobot (PLR) will make this much easier and enable you to generate normalisations that would normally not be possible, both in terms of throughput and complexity.

But, if your normalisations are relatively simple then I would recommend reconsidering whether such complexity is necessary or whether a simple single-channel system like the Opentrons OT-2 can do the job and save time :slight_smile:

1 Like
9

Thank you so much for all the information! The budget is around 100K and I was looking at potentially getting the STARlet.

3 Likes