Liquid Handling for Reproducible Science Post 2: Setting Up an In-House Verification System

November 16, 2017 by Thomas Rawlins Leave a Comment

In the second post of a 2 post series about liquid handling for reproducible science, we hope to provide some insights to help researchers to set-up a reliable and robust in-house liquid handling verification system.

Why set up verification methods for a liquid handling platform?

Liquid handling can be a significant source of variability when preparing samples for genetic analysis by qPCR or NGS, especially when miniaturizing workflow to conserve precious reagents and save on costs.  Accuracy and precision both play a critical role in the optimization of reagent and sample ratios when miniaturizing reaction volumes. Setting up a volume verification system in-house provides a cost-effective means to ensure that automated liquid handlers are mechanically sound and enabling verifiable science by transferring liquid accurately and precisely.  Extensive liquid handling verification can be a lengthy process, but at the bare minimum liquid handling performance should be validated for each liquid class at the volumes dispensed in a given application.  Note that only a single liquid class needs to be validated when working with a positive displacement instrument.

Measuring Accuracy Directly

Remember from our past blog post that accuracy and precision are calculated differently and can be measured in distinct ways.  Scientists can source commercially available volume verification systems that allow for both accuracy and precision to be assessed in a single measurement, such as the Artel MVS Multichannel Verification System.  However, if one aims to generate in-house methods for volume verification, the most pragmatic and efficient approach is to measure accuracy gravimetrically (direct method).  For example, researchers can calculate the average dispensed volume by taking the difference of the weight of an empty plate and the weight of a filled plate, dividing by the number of filled wells, and accounting for the density of the transferred liquid.  The systematic error, as discussed earlier, will be the difference between the nominal (target) volume and the measured volume.

Measuring Accuracy Indirectly

Note that when transferring smaller volumes of liquid, evaporation becomes a larger source of measurement error in gravimetric readings.  If transferring less than 5 µL of liquid, an absorbance based method can be utilized in order to assess the accuracy of your liquid handling platform.  In contrast to gravimetric methods, absorbance based approaches rely on comparison to standards of a known concentration.  A standard curve can be generated through a serial dilution of a known stock standard, and the absorbance reading of a target liquid transfer can be compared to this standard curve to indirectly measure the transfer volume.  Here, a highly reliable syringe or calibrated pipette must be used to generate the standard curve, and triplicate measurements are highly recommended. 

Measuring Precision Indirectly

Gravimetric approaches to measuring random error are not a viable alternative due to the number of measurements that need to be taken in order to obtain a statistically significant estimate of precision.  To have a statistically sound in-house verification system to measure random error, one will need a standard calibrated plate reader capable of reading at the absorbance maximum of a dye of choice.  Fluorescein, as one of the most commonly used dyes, can be a convenient selection.  It is important to calibrate measurement equipment separately from the liquid handling platform in accordance with the manufacturer's recommendations. 

Lastly, this approach requires solutions that mimic the viscosity of each of your liquid classes.  For example, fluorescein can be mixed with 25% glycerol to mimic a standard mastermix.   Again, when working with a positive displacement instrument, viscosity has virtually no impact on transferred volume allowing for the validation of volumetric performance of a liquid handling platform through a single target liquid.

What does the workflow look like?

Setting up an in-house system can take some time to get into place.  However, once a methodology has been established, verification of a liquid handling platform can be performed in minutes.

The following example outlining the measurement of random error as an estimation of precision of a 10 µL transfer of 25% glycerol to mimic mastermix provides a starting place for researchers to attempt an in-house verification test:

  1. Make a concentrated master solution of your chosen dye in water or suitable buffer
  2. Make a serial dilution of the dye master in a suitable microplate using the highest recommend working volume for that plate type (200 µl for a typical 96-well flat bottom plate)
  3. Read the plate at the wavelength corresponding to the dye absorption maximum
  4. Identify the dye dilution yielding the readings closest to 1 Au 
  5. Calculate the dilution of the dye necessary for obtaining 1 Au after diluting 10 µl of the dye solution to 200 µl
  6. Make a test solution containing the dye dilution from step 5 in 25% glycerol
  7. Dispense a test plate with 10 µl of the test solution and 190 µl of suitable diluent. Note that for aqueous solutions including a surfactant can greatly reduce the appearance of the meniscus, and associated readout error. Two suitable examples are TERGITOL™Type NP-7 and more common Tween-20.
  8. Centrifuge the plate, mix via orbital shaking, centrifuge again
  9. Read the sample plate at the wavelength corresponding to the dye adsorption maximum
  10. Import the data to your favorite spreadsheet/data analysis software and calculate the CV using the formula suitable for instrumental setup/assay.

Single Dispense vs Multi-Dispense:

Many users try to get around costs of consumable tips within their workflow.  It is important to verify liquid handling systems in accordance with the intended use case.  Note that different types of liquid transfers result in variations in the transfer volume:

  • Single-Dispense = Using a single tip to aspirate and dispense once.
  • Multi-Dispense = Using a single tip to aspirate 3 times the target volume and then dispense the target volume 3 times.

Liquid handlers transfer liquids most accurately and precisely when performing a single-dispense, as defined above.

Conclusion:

Verifying the performance of a liquid handling platform is critical to the generation of trustworthy data.  Further, performing verification tests in-house provides organizations with the utmost visibility to sources of variability while minimizing costs.  Compared to commercially available verification systems, researchers can save tens to hundreds of thousands of dollars a year when implementing an in-house test.

Click here to schedule a liquid handling consultation

Posted in: Liquid Handling, Mantis, Tempest

Subscribe to Email Updates

Recent Posts