The pride of the leselab is our pipetting robot Biomek FXp from Beckman Coulter, which arrived earlier this year. Together with a team from Beckman Coulter we are currently setting it up for our high trough put sequencing and sample processing pipelines.

Figure 1: First look at the Biomek FXp with a bit of labware on the deck to test the first protocols.

The first big projects we will carry out on the Biomek will be DNA extraction of single invertebrate samples with amplification of the DNA barcoding region. We will do this for 2.000+ individual specimens, which were identified by Finnish environmental offices for stream quality assessments, as part of a proficiency testing by the Finnish Environmental Institute (SYKE, cooperation with Kristian Meissner). Basically, we are using DNA to check if the different offices have identified the larvae specimens correctly, because only if species are identified correctly they can be used as indicators for stream health.

Figure 2: The first invertebrate samples from Finnland have already arrived.

Processing samples for DNA barcoding is in principle quite simple: Extract DNA of individual specimens, amplify the DNA barcoding region (COI Folmer) with PCR, clean up the reaction with EcoI / PastAP, add sequencing primer and send the samples away to a company for Sanger sequencing. All of this can be scaled up to a 96 well plate format, to process as manny samples at once as possible. This however gets tricky, if you start to get drop outs of individual samples in extraction, PCR, and Sanger sequencing! While the reactions are set up quickly manually, the necessary repeats and collecting samples from different wells because of dropouts is where things become tricky and are prone to errors when doing it manually.

So the task is now to set up a workflow for the robot, that can handle dropouts on several levels with minimal human intervention. I was ask by the Beckman Coulter team to give an overview what we had in mind, and i thought I share it here, as it might be interesting for others as well.

Steps needed for DNA barcoding with a robot

Here is a quick overview of the needed steps which have to carried out with “roboter friendly” reagents and laboratory methods. As in between human interaction is necessary (suppling new reagents to the robot, centrifuge plates after incubation etc) it makes sense to not have the whole protocol run as one big routine, but divide it into different parts.

The robot is supplied with a 96 well plate containing tissue pieces of each specimen.

1.) DNA is extracted with the AGENCOURT DNAdvance – Genomic DNA Isolation Kit
2.) DNA concentrations are quantified with the Fragment Analyser (Plate for measuring is prepared by the robot) Extraction plate
3.) A new 96 well plate is prepared, where DNA is diluted to a concentration of e.g. 10 ng/ul (using the FA concentration data) and failed DNA extractions are filled up with ones that worked. DNA Plate
4.) A PCR plate is prepared by the robot (using a PCR master mix) and 1 ul DNA from the DNA plate added. PCR is carried out in the thermo cycler in the pipetting robot. Then centrifuged down.
5.) PCR success is checked on the Fragment Analyser (Plate again prepared by the robot).
6.) PCR reactions which worked are transferred to a Cleanup Plate, to remove dntps and primers prior to Sanger sequencing.
7.) 5 ul of cleaned sample are transferred to the sequencing plate which contains 5 ul of 10 pmol/ul primer (only forward or reverse direction). Sequencing plate (Plate A: Forward, Plate B: Reverse and optional)

The big challenge is to keep track of all the samples. One “master” csv table might be most convenient so the robot can write down where in wich plates are samples and also to automatically replace samples which did not work so there is always a filled up 96 well plate used in the next steps. Additionally, DNA concentrations, and PCR / Sanger sequencing success has to be feed back to the robot. Also, how often a PCR / sanger sequencing has failed for a given sample, so that each sample is only repeated e.g. 3 times and then ignored, so problematic samples are not repeated over an over.

Additionally, protocols should be set up very modular, as for example we might not always measure DNA concentration oder PCR success whit samples that work really well (running samples on the Fragment Analyser almost gets more expensive than Sanger sequencing it self).

Figure 3:
Figure 3: What sounds simple gets quite complicated quickly. Overview of the different steps needed. Each step is planned as an individual program, as there is hands on time on the robotor needed any way to supply concentration data, supplies and reagents.

Step 1: DNA extraction plate!

In the following I describe necessary steps. Every thing is done by the robot, except for “hands on” bullet points

  • Hands on: Supply extraction solution in 20 ml falcon (Protinase K, DDT, Lysis Buffer – shake!).
  • Pipett 200 ul into each well of the extraction plate (Reuse the same tips, clean tips, store in “extraction plate tip box”)
  • Hands on: Add tissue in each of the well, supply the robot with csv indicating sample IDs and positions, incubate plate over night (on robot, and thermo shaker when doing several plates), centrifuge plate down.
    • Potentially shorten incubation time to 3 hours!
  • Carry out extractions (potentially have this as separate protocol, so the robot can be used over night).

Step A: Measuring concentrations – A more general protocol?

To enhance PCR success measuring and adjusting DNA concentrations might be necessary (using the Fragment Analyser, FA). This step however might be optional, as running samples on the FA can be costly and not necessary if DNA is of good quality.
The process for measuring the DNA extraction success is more or less identical to the detection of PCR success with the FA (the kit for PCR products might need a well where a ladder is added). Other reagents are used, but the pipetting the robot has to do is the same, thus this step can be used to check PCR success as well.
Also there should be an option to specify how many 12 well rows should be filled, as the protocol is something that might be used frequently. Also other tasks as exchanging the storage solution could be useful? Maybe 3D printing / building a insert for the FA bottles + liquid level sensing tips? Add a graphical interface (could even be tracking wells available etc)!

  • Hands on: Supply robot with fresh 96 well plate and FA buffer (Freeze robot suited aliquots for 96 well plates)
  • Pipett 22 ul FA buffer into plate and add 2 ul of each DNA extraction (96 head, clean tips, they could be reused in follow up cleanups – if it is possible to clean them DNA free!?).
  • Potentially mix plate by pipetting up an down!? – No centrifuging down required as air bubble free.
  • Hands on: Prepare FA, place plate for measuring in FA and start measurement.

Step 2: Prepare DNA plate

If a DNA extraction success and concentration was assessed with the FA, DNA concentration of samples should be adjusted to e.g. 10 ng/ul and samples which did not work should be replaced with ones that worked on the new DNA plate

  • Hands on: Supply FA with csv table containing DNA concentrations, supply water and fresh plate
    • Some samples might have a concentration below 10 ng/ul
  • Pipett water and DNA into wells. Keep the cleaned tip box, as it can be used for PCR and Sanger plate set up etc. Use the same tip for the same samples. Make sure to place the tip box on the robot in the right orientation!

Step 3: PCR

  • Hands on: supply PCR reagents, clean plate for PCR.
  • Set up and run PCR.
  • Hands on: Return reagents to the freezer, spin down PCR plate after PCR has run.

Step A: Asses PCR success

Same as above, but now with the PCR plate. Requires hands on time for the FA setup.

Step 4: ExoI FastAP cleanup and sequencing plates

Before PCR products can be cleaned up, primers and remaining DNTPs have to be removed from the PCR products. Also, it has to be decided here if the sequencing should be done for one or both directions. This directly influences the amount of needed ExoI / FastAP reagents.

  • Hands on: Supply ExoI / FastAP and clean plates / sequencing primers (100p).
  • Mix Exo / Fast AP and distribute 0.75 or 1.5 onto 96 well plate (depending on if 1 or 2 directions should be sequenced).
  • Incubate plate at 37°C for 15 minutes (don’t hat lid), and 95°C at for 25 minutes (with heated lit) – Potentially not necessary to spin down if lid heating and cooling sample down does not lead to condensation on top?
  • Make primer dilutions while incubating!
  • Hands on: Spin plate down (maybe).
  • If only one direction is sequenced, add 5 ul 10p primer to plate
  • if both directions should be sequenced, pipett 5 ul cleaned up PCR product into fresh plate and add diluted primers accordingly.
  • Hands on: Seal plates with tapes and add sequencing barcode
  • Hands on: Assemble sequences and report back to the robot which did not work and have to be repeated.


Developing a complete pipeline for pipetting robots can be actually quite complicated! Things might be optimized once we gain more experience, but at the moment it looks like still some hands on time is necessary. But as long as the robot is able to handle sample tracking and replace and repeat extractions / PCRs that did not work, it is still a good alternative to manually barcoding 2.000 samples.

Using the Biomek FXp for DNA barcoding!

Leave a Reply

Your email address will not be published. Required fields are marked *