384 Well Pipetting Precision and Accuracy Testing Using the Sciclone Advanced Liquid Handler
Larry Simonson and Lynn Jordan
Caliper Life Sciences Inc, Applications Laboratory, Hopkinton, Massachusetts, USA
Initially, work in 96 well microplates was commonly carried out using manual pipetting. As the demand for greater throughput of plate processing has grown, automated liquid handling has become a necessity for most laboratories. The most critical aspect of an automated liquid handler is to understand its performance, typically determined by precision and accuracy. At Caliper Life Sciences, precision and accuracy testing is achieved by dispensing a tartrazine solution into microplates containing water. This is referred to as a "liquid-to-liquid" dispense. The purpose of this paper is to report the performance obtained by dispensing 0.5 to 2μL dimethyl sulfoxide (DMSO) samples containing tartrazine into 384 well plates containing water. Precision and accuracy results are shown for this work.
Materials and Methods
Testing was performed using the Sciclone Advanced Liquid Handler (ALH) and a Bulk Reagent Dispenser (Sciclone pipettor option). Absorbance was measured at 450nm using a Wallac Victor2 Multilabel Counter (PerkinElmer).
Tartrazine (Sigma Chemical Company, Catalogue number T-0388) is dissolved in Dimethyl Sufloxide (J.T.Baker, 'PHOTREX' Reagent). A stock dispensing solution was prepared by dissolving 1.5g of tartrazine into 1L DMSO. The stock solution is for dispensing volumes of 0.5μL and 1μL. The concentration is reduced as the sample volume is increased to keep the absorbance measurement within the linear dynamic range of the detector. For the 2μL dispense testing, a 1:2 dilution is carried out on the stock solution. The tartrazine solution is dispensed into plates containing a volume of 50μL of degassed deionized water. The deionized water is degassed by bubbling helium through the water for a period of 15 minutes.
Qualification of a Reagent Dispenser
In our testing for a "liquid to liquid dispense", a Bulk Reagent Dispenser on a Sciclone Liquid Handler is used to dispense 50μL of degassed water into the plates. Qualification of the accuracy of this dispense is performed gravimetrically.
Calibration of the Plate Reader
To calibrate the plate reader, 75μL of degassed deionized water is pipetted into a 384 well polystyrene plate (Nunc #265196). To this plate, 1μL of the Tartrazine solution is manually added to each of 10 wells by successive 1μL dispensing from a filled 10μL Hamilton micro syringe. After manually mixing with multiple aspirate-dispense cycles using a hand pipetter, the wells were read for absorbance reader. Similar procedures were done for the 0.5μL and 2μL plates. The average absorbance reading for these manual standards was used in a single point calibration routine. This assumes a proportional relationship between the volume of dye in the well and the well reading.
Setting up the Sciclone
In the Sciclone software, a layout is created with two reservoirs, (Nunc #242811), and five 384 Well, clear, non-treated plates (Nunc #265196). Prior to running the Sciclone protocols, the accuracy of the Sciclone is verified. Using the values measured above in Calibration of the Plate Reader, the absorbance values obtained are compared to those obtained using the Sciclone dispense. To achieve the desired accuracy, under the "Liquids" tab, a Liquid called "LDMSO" is created and given a first order calibration value of 0.94. This same value (0.94) is used with all three volumes described in this note.
Running the Sciclone Protocols
The Sciclone is programmed to aspirate a volume of tartrazine solution. After dipping the cannulas into a rinse solution to remove any residual external tartrazine solution, they are dispensed into a clear bottom plate containing water for the "liquid to liquid" dispense. The procedure dispensed into 5 plates sequentially. Apart from appropriate changes in volumes, the same protocol is used for the 0.5μL and 2μL tests.
Below is an example of a Sciclone protocol for a 1.0μL dispense:
Mixing the Plates
After pipetting the 5 plates, an extra 25μL of DI water was added, again using a separate qualified reagent dispenser as described in MATERIALS AND METHODS to aid in the mixing of dissimilar liquids in the well. Next, the plates were given 15 shake cycles on a Molecular Devices V-max reader and allowed to set for 15 to 30 minutes before they were read.
Reading the Microplates
Plates were read for absorbance at 450nm. The raw results were exported to ExcelTM and CVs were calculated for each volume.
The three tables to the right show the precision and accuracy observed in this testing. The accuracy is shown as the average volume dispensed. The precision is given in CV for the individual plates and the 5 plate average at each volume. The values in the tables were imported from Excel sheets.
The experimental data demonstrates precise pipetting performance that can be attained using the Sciclone ALH. While Factory Release Dispensing Specifications for the Low Volume Head are 8% CV for volumes of 2μL and below, and 3% CV for volumes of 2 to 25μL, in this work we achieved CV's that are significantly better than the specifications. Interestingly, the first plate dispensing 0.5μL generally tends to show a higher volume and less precision than the remaining 4 plates in the dispense sequence. This effect is believed to be due to any residual tartrazine solution on the external surfaces of the cannula. These contributing factors can create a noticeable effect due to the higher concentration of the tartrazine solution for the lowest dispensing volumes. However, taking the result of the first plate into account, these results are still within in the effective working range for most applications.