A high-throughput quality control method for assessing the serial dilution performance of dose – response plates with acoustic ejection mass spectrometry

This study aimed to develop a streamlined method for evaluating the dilution ratio of drug dose – response plates created by automated liquid handlers in the early stages of drug discovery. The quantitative techniques commonly used for this purpose have restrictions due to their limited linear dynamic range and inaccuracies in assessing serial dilution performance. To address this challenge, we describe a method based on acoustic ejection mass spectrometry (AEMS). The method involves using standard compounds and an internal standard to evaluate each dilution point in quality control (QC) plates. The samples are transferred to a chromatography-free tandem mass spectrometry system through an acoustic source, enabling the analysis of one sample per three seconds from a microtiter plate. This approach provides precise, accurate, label-free, and rapid data acquisition to support high-throughput screening efforts.


Introduction
The development of high-throughput and ultra-high-throughput screening platforms has enabled rapid screening of hundreds of thousands of compounds against biological targets.This is achieved through highly automated and miniaturized assay formats based primarily on microplate technology and requiring liquid handling techniques suitable for microliter and sub-microliter volumes of samples and reagents [1][2][3][4].Compound potency is often measured in a concentration-dependent manner that requires dose-response (DR) plates [5,6].Serial dilution techniques enable a wide range of compound concentrations to be tested in a single experiment, thereby identifying the most effective concentration range and determining the potency of a compound.Serial dilution is a cost-effective approach that also reduces the amount of compound required for testing.
Automated liquid handling systems have played a critical role in laboratory applications for bio-pharmaceutical product development by facilitating tasks that require precise and accurate pipetting, including sample preparation, serial dilution, and reagent transfer.Acoustic droplet ejection (ADE) is a cutting-edge technology for creating DR plates in 384-well or 1536-well format through a direct dilution approach, allowing the transfer of various volumes from a constant concentration.The process involves ejecting droplets from a source plate into an upside-down destination plate, with both plates being held in designated plate holders within the liquid handler.Echo acoustic liquid handlers are now widely used because of their fully automated, noncontact design and the highly reproducible nature of their operation.This ensures minimal variability between replicates and makes the devices ideal for generating dose-response plates [5][6][7][8][9][10][11].
The accuracy of serial dilution methods is critical for generating accurate dose-response data.The most common methods used to evaluate dilution ratios are gravimetric, fluorometric, and photometric methods or a combination of these methods [1][2][3][4][12][13][14][15][16].Gravimetric methods measure the mass of the sample and provide accurate and precise measurements.However, this approach can be time-consuming and takes effort to provide precision measurements for individual channels or wells on a microtiter plate [12,13,15].Fluorometric methods use fluorescent dyes to measure the concentration of the sample, and such approaches can be highly sensitive and specific, but they may be affected by the quantitative range of fluorescein signals produced by the fluorescence reader [1,2].Photometric methods use spectrophotometers to measure the absorbance or transmission of light by the sample.They are relatively simple and fast but may be affected by variations in dynamic ranges and saturation points [13,16].Each method has its own advantages and disadvantages, and the choice of method depends on the requirements of a specific experiment.
Acoustic ejection mass spectrometry (AEMS) (commercially marketed as Echo MS) is a novel technology that offers several advantages over traditional methods [17,18].First, it is a label-free approach, which eliminates the need for markers or labels to detect samples or compounds.This is especially beneficial because labeling can be time-consuming and costly.Additionally, AEMS is designed to be a non-contact technology, which minimizes the risk of cross-contamination and sample carryover between experiments, a critical consideration in high-throughput screening and drug discovery.Finally, the AEMS system is highly efficient, being capable of dispensing up to three samples per second, enabling fast and effective processing of large sample sets.These label-free, non-contact, and high-speed capabilities make AEMS an attractive choice for quantitative measurement-based sciences, including high-throughput screening, quantitative measurement, and biomarker analysis [19][20][21].AEMS is also being applied in situ enzyme kinetics studies; in absorption, distribution, metabolism, and elimination (ADME) assays; in pharmacokinetics studies; and in parallel medicinal chemistry experiments [22][23][24][25][26].
Here we describe a high-throughput quality control (QC) method to assess the performance of automated liquid handling systems.By using AEMS, a robust and efficient automation process was established for accurate data acquisition and processing in a 384-well format.To the best of our knowledge, it is the initial assessment method utilizing a mass spectrometer to evaluate the dilution performance of DR plates.

Chemicals
Methanol and acetonitrile (LCMS grade) were acquired from Fisher Scientific (Waltham, MA).Ammonium fluoride, verapamil, and dimethyl sulfoxide (DMSO) were purchased from Sigma-Aldrich (Burlington, MA).Carbamazepine was ordered from Biosynth (Gardner, MA).Warfarin was obtained from Fluka Chemical Corp.(Ronkonkomak, NY).Milli-Q water (Millipore, Molsheim Cedex, France) was used in the reagent.Echo-MS qualified 384-well polypropylene (PP) and cyclic olefin copolymer (COC) microplates were purchased from Beckman Coulter Life Sciences (San Jose, CA; cat.nos.C74290 and LP-0200, respectively).PP plates were suitable for both the Echo-based plate replication process and the AEMS analysis, whereas COC plates were only compatible with the Echo liquid handler.
Carbamazepine, verapamil, and warfarin are commonly used standards for in vitro ADME assays, particularly in cell-based permeability studies.Each of these compounds has unique characteristics.Verapamil, a p-glycoprotein transporter (P-gp) inhibitor and calcium channel blocker used to treat hypertension and angina, is stable in storage, safe to use, and highly soluble in water [27][28][29].Carbamazepine, which is not a P-gp substrate [27], is relatively safe, with a long half-life and good storage stability.It is moderately soluble in water and is commonly used as a standard for analytical methods in pharmaceutical research.Warfarin is an anticoagulant medication used to prevent blood clots.It is stable in storage, moderately soluble in water, and has been used as an internal standard in high-throughput MS screening [28].

AE liquid dispenser system
An Echo 650 Acoustic Liquid Handler (Beckman, San Jose, CA), was used to dispense compound solutions in 384-well plates and to prepare Fig. 1.Echo MS dilution plate workflow.(1) Dose-response plates were made using an Echo 650 system with 10-point dilution points, using two layouts: an alternating (even and odd) dilution series to mimic liquid handler formating and a blocked series.(2) Fifty nanoliters were then transferred from the Echo MS dose-response plates to new 384-well PP assay plates for Echo MS analysis.(3) Cosolvent containing 400 ng/mL of warfarin was then added to each assay plate.For the cosolvent delivery study, two systems were used: an Apricot system and a BioTek MultiFlo FX system.(4) After centrifugation for 5 min, the plate was analyzed by Echo MS at two different volumes, 2.5 nL and 20 nL.(5) For this study, raw data files were processed in the Sciex OS-MQ software.(6) Statistical analysis and plotting were performed in R/RStudio.dose-response plates (Fig. 1).This instrument uses acoustic droplet ejection technology, enabling contactless transfer of DMSO through focused acoustic energy.By monitoring the acoustic energy reflected from the fluids in each well of a plate, the instrument can determine the hydration level of the DMSO and the depth of the well fluid.The system facilitated the direct transfer of compound droplets in precise 2.5-nL increments from a source plate to a destination assay plate in a 384well format.This eliminated the need for intermediate dilutions of active compounds, thereby streamlining the process.

Solvent dispenser systems
A MultiFlo FX multimode dispenser (BioTek, Taunton, MA) and an Apricot S2 pipette (Apricot Designs, Covina, CA) were employed to dispense the cosolvent (30 % acetonitrile in Milli-Q water with 400 ng/ mL warfarin) at low to medium volumes.The MultiFlo FX has eight pipette tips and can dispense volumes ranging from 50 nL to 50 μL into multiple wells of a microplate simultaneously.For this experiment, cassettes with 1-μL and 10-μL volumes were used to dispense the cosolvent.The system uses a peristaltic mode with a series of rollers that compress a flexible tube, creating a positive displacement that moves the liquid through the tubing and out of the cassette.The user-friendly interface enables programming of the desired volume, well location, and dispensing speed, along with continuous dispensing and priming.The Apricot S2 pipette is an electronic air-displacement pipette that dispenses accurate volumes of liquids in a 96-or 384-well format.Its motor-driven piston aspirates and dispenses liquids through a disposable tip.The desired volume is selected via the digital display and buttons on the pipette, and the motor-driven piston draws the liquid into the disposable tip.The pipette can be programmed for repeat dispensing and is comfortable to use for extended periods.
To set up the experiment, 50 μL of cosolvent was dispensed from the three delivery systems.Two plates were prepared for each system, resulting in a total of six plates for the experiment.The amount of cosolvent was chosen to ensure that there would be sufficient material for accurate analysis while avoiding the issues that could arise from using excessive cosolvent.

Preparation of stock solutions, calibration standards, and quality control samples
Two distinct stock solutions of carbamazepine and verapamil were prepared at a concentration of 10 mM in DMSO.These solutions were used for bulk spiking of calibration standards and quality control samples in method validation exercises, as well as for analysis of dilution ratios.The stock solutions were stored at room temperature.
For precision and stability experiments, 50 nL of carbamazepine and/or verapamil stocks were dispensed into each well of the 384-well destination plates, using the Echo 650 dispenser.Then, 50 μL of cosolvent (30 % acetonitrile in Milli-Q water with 400 ng/mL warfarin) was added either manually with a multi-channel pipettor or automatically with the MultiFlo FX system (with two different volume cassettes) or with an Apricot S2 pipette.
For dose-response plates, calibrators and QC working solutions of carbamazepine were prepared using the Echo 650 dispenser.Separate calibrator working solutions were created by dispensing the stock solution and backfilling with DMSO to make 3-fold dilutions, resulting in final concentrations of 10,000, 3333.3,1111.1, 370.4,123.5, 41.2, 13.7, 4.6, 1.5, and 0.5 µM.Similar QC working solutions were prepared with concentrations of 2000, 200, and 20 µM (Supp.Table 1).An assay plate with standards and QCs was then created by dispensing 50 nL of calibrator working solutions and QC working solutions with the Echo dispenser.Finally, 50 μL of cosolvent (30 % acetonitrile in Milli-Q water with 400 ng/mL warfarin) was added using the MultiFlo FX with a 10-μL dispensing cassette (Fig. 1).

Precision readout
Two different ejection modes were used to acquire carbamazepine, verapamil, and the warfarin internal standard with the Echo MS system.In the consecutive mode, 25 consecutive 2.5 nL transfers were ejected from a single well, with a total of 12 wells being used.Alternatively, the cycling mode was used to dispense 2.5 nL from each of the 12 wells for a total of 25 cycles.These ejection modes were used to evaluate the precision and reproducibility of the Echo MS system.

Stability of plates and analytes
To assess the stability of the analyte and the quantitative method, two 384-well plates were used.After 50 nL of carbamazepine was dispensed into each well of the plates by using the Echo dispenser, one plate was stored at room temperature and the other plate was stored at − 20 • C. Each day, 12 wells were evaluated, and 50 transfers were taken per well.The cosolvent was added manually, using a multichannel pipette.The study was carried out over a period of 8 weeks, resulting in a total of 4200 replicates per temperature.The replicates were acquired on days 1, 2, and 3, then at weeks 1, 2, 3, and 8.This approach enabled the stability of the samples under different storage conditions to be evaluated and provided a better understanding of the impact of storage temperature on the analysis results.

Preparation of dose-response plates for AEMS analysis
To prepare the DR plates, two plate layouts were used.The first layout involved direct dilutions in alternating odd-or even-numbered columns.For example, odd-numbered columns 1, 3, 5, 7, 9, 11, 13, 15, 17, and 19 or even-numbered columns 2, 4, 6, 8, 10, 12, 14, 16, 18, and 20.The second layout involved blocks of columns 1-10 or columns 11-20.Quality controls were consistently placed in columns 22, 23, and 24, with a DMSO blank in column 21.Three assay plates were prepared for each layout per day (intra-day), and two separate days of analysis (inter-day), which resulted in a total of 12 plates being analyzed.Two different ejection volumes were used: 2.5 nL for steps 1-10 and 20 nL for steps 6-10.

AEMS system
Before being loaded into the Echo MS system, the source microplates containing the sample cosolvent were centrifuged (at 537 RCF for 5 min) to eliminate gas bubbles and to ensure a uniform fluid meniscus shape.
To collect samples, the Sciex OS Analytics Software (version 2.1.6.59781)(Sciex, Concord, Ontario) was used to control an Echo MS system operating in multiple reaction monitoring (MRM) mode.The Echo MS system consisted of an externalized transducer assembly from an Echo MS autosampler, an open port interface (OPI) connected to a carrier solvent pump, and a transfer capillary that leads to a OptiFlow Turbo V Ion source of an AB Sciex Triple Quad 6500+ system.
The system used a carrier liquid consisting of methanol supplemented with 1 mM ammonium fluoride, flowing at a rate of 400 µL/min.This generated a stable vortex at the OPI inlet to optimize the signal performance.The contactless sampling method involved ejecting either 2.5 nL or 20 nL directly from the microtiter plate wells containing the samples into the carrier liquid vortex of the OPI.The ejection operation mode was set to "normal", defined as an interval of 1000 ms with a sampling frequency of less than 1 sample per second.Moreover, an additional delay of 2000 ms was introduced in the AE method.The electrospray ionization (ESI) source of the triple quadrupole MS instrument operated in positive ionization mode.The nebulizer gas (GS1) was set to 90 psi, and the heater gas (GS2) was set to 70 psi.The curtain gas and CAD (collision-activated dissociation) gas were set to 20 psi and 9 units, respectively.The chosen mass transitions for analysis were 237.1→194.1 for carbamazepine, 455.2→165.1 for verapamil, and 309.1→251.1 for warfarin.The following MS parameters were employed for analyte measurements: ion source temperature: 500 • C; spray voltage: 5500 V; dwell time: 15 ms; pause time: 5 ms; Q1 and Q3 operated at unit resolution; entrance potential (EP): 10 V. Analyte-specific settings, including declustering potential (DP), collision energy (CE), and collision cell exit potential (CXP), were applied for the Q1 to Q3 transition for carbamazepine (DP 80, CE 29, CXP 11), verapamil (DP 100, CE 38, CXP 9), and warfarin (DP 50, CE 29, CXP 8).After sample batches were acquired, data processing was conducted using the following signal integration parameters: minimum signal-to-noise (S/N) threshold of 2 and high Gaussian smoothing with moving average.The expected retention time was set to a minimum of 0.02 min.
2.9.Data analysis and software

Outliers
To measure the variability of the Echo MS transfers in the precision and stability studies, the median absolute deviation (MAD) of the internal standard (warfarin) was calculated for each well (n = 25 and n = 50, respectively), and outlier transfers were defined as those whose peak area exceeded 4.5 times the MAD away from the median area.To eliminate variation due to the Echo MS when evaluating the compound precision, any internal standard outliers were excluded from the calculations, and any well with an internal standard outlier frequency exceeding 10 % was also excluded.The MAD for the different cosolvent delivery methods and the dose-response plates was calculated based on the median of each individual plate (Eq.( 1)), and outliers were defined and excluded as described above.

Variation
The coefficient of variation (CV) was calculated for individual wells (for precision studies), overall, by temperature and by day (for stability studies), by plate and by system (for cosolvent delivery system studies), and for each compound concentration (for dose-response plates):

Linearity
To assess the linearity, two sets of standard curves were generated using 10 calibration standards, which covered a range from 0.5 nM to 10 μM with an ejection volume of 2.5 nL as well as a range from 0.5 to 41.2 nM with an ejection volume of 20 nL.The appropriate weighting factor for linear regression was set as 1/x 2 .

Internal standard QC
Before the dilution series was evaluated, the quality of the Echo MS runs was assessed.This involved calculating the CV and standard deviation of the internal standard (warfarin) for each plate, along with the MAD, and identifying outliers as described in Section 2.9.1.If the internal standard CV of a plate exceeded 10 % or the frequency of outliers was greater than 5 %, the plate was deemed to not meet the quality criteria for the MS run.In such cases, the plate underwent re-analysis to ensure accurate results.Any wells identified as internal standard outliers were excluded from the compound analysis because there may have been an issue with the ejection process for those particular wells.Additionally, any wells flagged as "bad ejections" by the Echo MS itself were also excluded from further analysis.

Normalization
To obtain dose-response normalized concentrations, the area of the compound was normalized to the highest concentration (10 µM).This involved rescaling the areas of the 20-nL transfers to match those of the corresponding 2.5-nL transfers.By performing this normalization, the compound concentrations across the dose-response range were adjusted to a consistent reference point, facilitating accurate comparisons and analysis.

Dose-response plate pass/fail criteria
Acceptable criteria for a serial dilution plate were determined by precision studies and by running multiple dose-response plates to establish the baseline behavior of the Echo MS and Echo 650 in creating 3-fold dilution series.In the final evaluation of the dilution ratios, the 2.5-nL ejections were used for the ratios of the first five dilutions, and the 20-nL ejections were used for the ratios of the last four dilutions.The variation of the ratio of the last two steps was 25 % higher on average than that of the previous ratio, even with 20-nL ejection, and this ratio was, therefore, excluded.The MAD of each compound concentration for each plate and ejection volume was calculated, and an outlier was defined as any well that was 4.5 times the MAD from the median.The overall ratio and the CV of the ratios were calculated both with and without the compound outliers.

Software
The Echo MS data files were processed using Sciex OS-MQ Analytics Software (version 2.1.6.59781).The processed data were subsequently exported and analyzed in R, using RStudio IDE [30,31] and ggplot2 [32] for visualization.The R code for analyzing the goodness of a serial dilution plate, using the Echo MS data and producing the table seen in Table 3, can be found at https://github.com/maryashleyrimmer/SerialDilutionsReport.

Precision
The precision analysis of the carbamazepine, verapamil, and warfarin internal standard revealed that consecutive sampling from a single well resulted in greater variation than was observed with the cycling transfers mode (Fig. 2).However, the variation for both analytes and warfarin largely remained below 10 %.These observations are consistent with previous reports on the performance of the Echo MS [18,20].The CV for the analyte peak area was less than 10 %.Outliers were identified separately for each compound, and any outliers detected were removed from the dataset.The analysis revealed five outliers (1.7 % of the total) in the consecutive mode, whereas only one outlier was observed in the cycling transfers mode.The high precision demonstrated by the cycling mode indicated that it was not necessary to have replicates from each plate/well and that a single injection into the Echo MS could provide reliable data points.Based on these findings, an acceptable error limit of 10 % was chosen as the quality criterion for the MS run.These results indicate that the ejection modes employed in this study were reproducible and generated consistent results with acceptable error rates.

Stability
A total of 8400 data points were generated during the 8-week stability test, in which two storage conditions were evaluated.Among these data points, only 81 (less than 1 % of the total) were identified as outliers and subsequently removed.Of the 168 wells analyzed, three had more than 10 % outliers, resulting in the exclusion of all 50 transfers from those wells.Specifically, two of these wells were from the room temperature plate and one was from the − 20 • C plate, accounting for less than 2 % of the total wells.Consequently, a total of 204 data points were removed, representing 2.4 % of the overall dataset.
After the outliers were removed, the absolute intensity of carbamazepine and warfarin remained consistent across all test points and storage conditions (Supp.Table 2).The overall variation for carbamazepine was less than 9 %, whereas that for warfarin ranged from 9.01 % to 10.73 % (Supp.Table 3).The consistent intensity of carbamazepine is indicative of the high precision of the liquid dispenser, the stability of the compound, and the excellent reproducibility of the Echo MS system.
The variations observed in consecutive sample ejections for each well, time point, and storage condition were consistent with the findings from the precision experiment, with values being close to 10 % (Fig. 3).Although the delay between each ejection into the Echo MS was set to 2 s, the variation occurred primarily in a few problematic wells (Fig. 3B).The use of the peak area ratio of the analyte and the internal standard (Fig. 3A) eliminated this variation, suggesting there was a correlation between the variation and the ejection volume.The change in the fluid meniscus shape in a specific well during the 50 consecutive transfers may have contributed to inaccuracies in the ejection volume.However, when considering the peak area ratio, the variation was effectively resolved.
The variation in the peak area ratio of carbamazepine and the internal standard showed a slight increase, from approximately 9 % to 11.5 %, for both storage conditions when compared to the variation in carbamazepine or the internal standard alone (Supp.Table 3).The variation of each analyte was accumulated to contribute to the overall higher variation in the peak area ratio.Furthermore, adding a fresh internal standard on each acquisition day had no significant impact on the results.
This study demonstrated that there was no significant difference in the stability of carbamazepine under the two storage conditions (− 20 • C and room temperature) over the 8-week period.Therefore, storing carbamazepine at − 20 • C or room temperature for up to 8 weeks does not lead to significant degradation.

Cosolvent delivery system
We evaluated the variability of three cosolvent delivery systems, namely the Apricot S2 system and the MultiFlo FX system with two different cassette volumes (1 μL and 10 μL) (Table 1).A total of 768 data points were analyzed for each system, and the CV was found to range from 7 % to 9 %.Additionally, the variation was assessed across plates, with two plates being evaluated for each system.The CV for these plates ranged from 3 % to 5 %.The number of outliers was also recorded: The Apricot system had five outliers, whereas the MultiFlo FX had three outliers with a cassette volume of 1 µL and two outliers with a cassette volume of 10 µL.These findings suggest that the three delivery systems  have comparable levels of variation.Overall, the data indicate that all three systems perform with high reliability and consistency.However, one should consider practical factors when selecting a delivery system.For example, the Apricot requires pipette tips, which can be expensive and wasteful, whereas the MultiFlo FX cassette is reusable.These practical considerations can make a significant difference, especially in high-throughput applications in which large numbers of samples are processed.Ultimately, the choice of delivery system depends on the specific needs of the user.In this study the MultiFlo FX was found to be the most efficient and cost-effective system.

Linearity
The standard curves, constructed using ten standards (ranging from 0.5 nM to 10 μM with an ejection volume of 2.5 nL) and five standards (ranging from 0.5 nM to 41.2 nM), exhibited a strong regression relationship.These curves were generated by plotting the peak area ratio of carbamazepine to warfarin against the quantitative concentrations (Supp.Fig. 1).The coefficients of determination (R 2 ) for both sets of standards were 0.99, with a sample size of 32 for each standard.
Additionally, we established a linear relationship between the Echo-MS responses and the volumes of the sample ejected within the range of 2.5 to 20 nL (Supp.Fig. 2).In this case, the sample size was 16 for each standard from steps 6 to 10.

Dilution ratio for dose-response plates
The dilution performance of the assay plates was demonstrated by plotting the dilution series of each row from one of the blocked plates (Fig. 4A and Supp.Fig. 3).All data points were standardized to the top concentration of 10 μM.In Fig. 4A, the 2.5-nL transfers are indicated by red points, and the 20-nL transfers are indicated by blue points.The triangles represent single-point QC points.The plot demonstrates the reproducibility of the Echo MS, with each series exhibiting a similar trend and clear differentiation between concentrations.
During the evaluation of the dilution ratio of DR plates, two ejection volumes, 2.5 nL and 20 nL per sample, were tested.The data analysis revealed excellent reproducibility and precision for the 2.5-nL transfers in the first eight steps, but a significant increase in variability was observed with the lower concentrations in steps 9 and 10 (Fig. 4B).However, using 20-nL transfers for the last five steps resulted in a notable reduction in variability, as shown by the blue boxplots.To calculate the overall dilution ratio for each plate, the first five steps from the 2.5-nL ejections and steps 6-9 from the 20-nL ejections were considered.These findings indicate that employing larger ejection volumes for low-concentration samples can enhance sensitivity and reduce variability, leading to more reliable results.
By employing the established outlier percentage and acceptable precision as a reference, the internal standard can be used to assess the mass spectrometry run for each QC plate (Supp.Fig. 4).This analysis enables the identification of any outliers or issues that deviate from the normal behavior of the Echo MS.If the mass spectrometry response of the internal standards exhibited uniformity across the entire plate, it served as an indication that both the cosolvent addition step and the Echo MS system were operating effectively.In cases where consistency was not achieved, QC procedures were required to be repeated with a newly generated assay plate.This enables differentiation between significant variations beyond the acceptable criteria for the dilution ratios, whether they are attributed to the Echo MS analysis or to the liquid handling protocol itself.
In this study, the overall dilution factors closely matched the theoretical value of 3 for both the alternating and block plates (Table 2).No difference was observed when the testing was performed in the inter-day and intra-day experiments.Our standard dose-response plates were created using an Echo 650 liquid dispenser, with each step being generated directly from the stock solutions.This approach eliminated the possibility of errors accumulating among the serial dilutions, in contrast to the aforementioned serial dilution strategy using other liquid  dispenser systems.The comprehensive dilution ratios across all 12 plates demonstrates the Echo system's capability to produce precise measurements and accurately prepare direct dilution series for DR plates, whether in alternating or block formats.For applications of methodology, actual dose-response plates can be generated using liquid dispensers with disposable tips, acoustic droplet ejection, or other suitable techniques.The concentrations and dilution ratios of these DR plates can then be assessed using our developed method.There are three possible assessment approaches.Control Compound in One Row: One option is to include the control compound (e.g., carbamazepine) in a single row of each dose-response plate.This approach offers the advantage of assessing all DR plates but has the limitation of only evaluating the specific row with the control compound.Full QC Plates with Control Compound: Alternatively, QC plates can be filled entirely with the control compound.This approach evaluates the overall performance of the liquid dispenser system during the generation of DR plates.However, it allows for periodic evaluations (daily or weekly) rather than continuous monitoring.The approach also focuses solely on assessing the procedure rather than evaluating individual DR plates.Monitoring Any Interested Compound: Given that the assessing method relies on mass spectrometry, concentrations and dilution ratios of any compound of interest applied in the DR plate can be monitored and evaluated if the multiple reaction monitoring (MRM) method has been developed.This represents a more advanced approach compared to the current reference QC method(s), which typically rely on fixed fluorescence or absorbance probes.

Conclusion
This study developed quantitative methods for assessing dilution performance, using carbamazepine and verapamil, with warfarin as an internal standard and with variation of less than 10 % in both cycling and consecutive shooting modes.The stability experiment confirmed the reliability of the analytical method, showing minimal outliers and similar stability for carbamazepine under different storage conditions.In the evaluation of cosolvent delivery systems, the BioTek MultiFlo FX system proved more efficient and cost-effective than the Apricot system, but the choice of system will ultimately depend on user-specific needs.These findings have important implications for the handling, storage, and monitoring of carbamazepine in laboratory settings.
Previously reported pass/fail criteria were used to assess the accuracy and precision of the serial dilutions for dose-response plates [33].Based on our study, QC plates will be deemed to pass if the outlier frequency is less than 5 %, the CV is less than 20 %, and the average dilution factor ranges from 2.80 to 3.20 (as shown in Table 3).Plates failing to meet these criteria will be classified as failed, thus ensuring the monitoring and performance of liquid handling systems to produce high-quality dose-response plates.
In this study, a new automated process using chromatography-free tandem mass spectrometry and an acoustic source from a microtiter plate was developed to evaluate QC plates.Using Echo MS, we established a reliable and efficient automation process for sample preparation, data acquisition, and processing in a 384-well format.This label-free method is cost-effective and highly accurate, providing reliable results and representing a significant breakthrough in highthroughput screening and drug discovery.

Fig. 2 .
Fig. 2. Precision of the Echo MS.The coefficient of variation for 25 ejections from each of 12 wells to measure precision of the Echo MS. (A) Each well was ejected into the Echo MS 25 times before the system moved to the next well.(B) All 12 wells were ejected once then cycled through 25 times.

Fig. 3 .
Fig. 3. Stability of carbamazepine.The peak areas of carbamazepine and warfarin (the internal standard) from each transfer were obtained on the Echo MS over a period of 8 weeks.Each well was analyzed 50 times, and the wells are denoted by color.Compound plates were stored at room temperature or − 20 • C. The coefficient of variation was calculated from all 600 transfers each day.(A) The area ratio of carbamazepine/warfarin. (B) The area of carbamazepine alone.

Fig. 4 .
Fig. 4. Dose-response plates.(A) Representative plots of the dilution series by row.Both even and odd dilution series are included for each row.Red points are 2.5-nL transfers and blue points are 20-nL transfers (rescaled), showing the consistency of the data scale regardless of the transferring volume.All data are normalized to the top concentration.Triangular points denote controls from each row, and the dashed gray line is the expected fit for a 3-fold dilution series.Plots for rows A, B, C, and D from a single plate are shown here.The complete set of dilution plots for this plate can be found in the supplementary material.(B) Boxplots showing the increased precision for the lower concentrations when using larger volume ejections into the Echo MS (2.5 nL vs. 20 nL).

Table 1
Internal standard variation by cosolvent delivery system.