A semi-automated ELISA workflow for SARS-CoV-2 Nucleocapsid IgG detection significantly improves throughput and reduces hands-on time Authors: Ruturaj R. Masvekar, Gabriela Mansano Do Nascimento, Aliyah Hannah A. Remoroza, Shweta Jain, Troy J. Kemp, Bruce K. Brown, Ligia A. Pinto Published: 2026 ABSTRACT Background:Enzyme-linked immunosorbent assay (ELISA) is widely used in clinical and research laboratories for applications such as disease surveillance, immunogenicity assessment, and vaccine evaluation. Traditional manual ELISA workflows, however, are labor-intensive, time-consuming, and susceptible to operator-dependent variability, which can limit throughput and reproducibility. Incorporating semi-automation into established protocols offers a potential strategy to improve efficiency while maintaining analytical performance.Methods:A validated manual ELISA protocol for SARS-CoV-2 Nucleocapsid IgG detection was adapted into a semi-automated workflow using an automated multichannel pipette, microplate washer-dispenser, and plate stacker to streamline key steps. Performance was compared across 10 manual and 30 semi-automated plates. Sixteen pre-characterized serum samples were analyzed to evaluate precision, accuracy, and overall assay validity, and internal controls were monitored to ensure compliance with predefined acceptance criteria.Results:The semi-automated workflow doubled assay throughput from 6 to 12 plates per analyst per day and reduced hands-on time by 75% (from 5.5 h to 1 h). Internal controls for all 40 plates met established acceptance ranges. Precision, measured as relative standard deviation, remained below 25% for all samples, and accuracy, expressed as percent error, remained below 20%. No loss of analytical performance was observed when transitioning from manual to semi-automated processing.Conclusion:Semi-automation markedly increases ELISA throughput and reduces hands-on time while maintaining precision, accuracy, and overall assay validity. These results underscore its feasibility as an efficient workflow enhancement with potential for broader application across diverse ELISA platforms pending further validation. Introduction Enzyme-linked immunosorbent assays (ELISA)s are invaluable for applications including, but not limited to, tracking disease outbreaks, detecting prior exposures, supporting vaccine development and immunogenicity evaluation, and informing public health interventions (Alhajj et al., 2025; Hornbeck, 2001; Porstmann and Kiessig, 1992). Despite its widespread utility, the conventional ELISA workflow remains highly manual, requiring multiple sequential steps such as reagent addition, incubation, and repeated microplate washing. These labor-intensive procedures not only limit throughput but also introduce variability due to operator-dependent factors (Alhajj et al., 2025). As the demand for large-scale testing and standardized immunogenicity assessments continue to grow, the limitations of manual ELISA workflows present a significant bottleneck. Automation technologies offer a promising solution by streamlining repetitive tasks, reducing human error, and improving assay reproducibility (Zhou et al., 2021; Zanchettin and Facciotti, 2022). Integrating tools such as automated multichannel pipettes, microplate washers and dispensers, and plate stackers into established protocols has the potential to substantially increase efficiency while maintaining analytical rigor. In this study, we developed and evaluated a semi-automated ELISA workflow for the detection of SARS-CoV-2 Nucleocapsid IgG antibodies. By incorporating automation into a previously validated conventional protocol, we aimed to determine whether throughput and efficiency could be enhanced without compromising assay validity, precision, or accuracy. Materials and methods Samples Serum samples from donors were obtained through three SeroNet Capacity Building Centers (CBCs): the University of Minnesota, the Icahn School of Medicine at Mount Sinai (Mount Sinai), and the Feinstein Institutes for Medical Research (Feinstein). All samples were collected under approved Institutional Review Board (IRB) protocols: STUDY00012418 (University of Minnesota), IRB#16–00791 (Mount Sinai), and IRB#20–1007 (Feinstein). The samples were heat inactivated, aliquoted and stored at −80°C until used. Critical reagents The semi-automated workflow was integrated into our pre-validated SARS-CoV-2 Nucleocapsid IgG ELISA protocol. For both workflows (conventional and semi-automated), identical lots of the following reagents were used: recombinant SARS-CoV-2 Nucleocapsid (NC) protein (SARS-CoV-2 N [1–419], N-terminal His6 tag; obtained from the Protein Expression Laboratory (PEL) at Frederick National Laboratory for Cancer Research [FNLCR]), internal standard calibrated to the WHO international standard (20/136), internal assay controls (Negative Control [CN], Positive Control–Low [CPL], and Positive Control–High [CPH]), goat anti-human IgG antibody-HRP conjugate (SeraCare; Cat# 5220–0330), and 2-component TMB substrate (SeraCare; Cat# 5120–0050). Instrumentation Depending on the workflow, the following pieces of equipment were used: multichannel pipettes (Rainin: Pipet-Lite XLS [manual] and E4 XLS+ [electronic]; Integra: VIAFLO), microplate washer and dispenser (Agilent; 406FX), microplate stacker (Agilent; BioStack), incubator (Thermo Fisher Scientific; Heratherm), and microplate reader (Molecular Devices; SpectraMax M3). Conventional ELISA workflow The conventional workflow for detecting anti-SARS-CoV-2 NC IgG was performed as previously described (Bullock Jr et al., 2024; Haynesworth et al., 2024; Hickey et al., 2023). NC antigen aliquots were thawed and diluted to a final concentration of 0.45 μg/mL in DPBS (Life Technologies; Cat# 14190–136), then used to coat Maxisorp 96-well plates (Thermo Fisher Scientific; Cat# 439454). Plates were sealed and stored at 2–8°C until use. Coated plates were used between days 2 and 7 post-coating. On the day of assay, plates were washed with PBS-T (0.05% Tween-20 in PBS) and blocked with freshly prepared blocking buffer (4% Skim Milk with 0.2% Tween20 in DPBS) for 90 min at 20°C. Standards (8-point, 2-fold dilution), controls (4-point, 3-fold dilution), and 16 pre-validated samples (8-point, 3-fold dilution) were prepared in 2 mL deep-well storage plates (Thermo Fisher Scientific; Cat# NC9792727) using blocking buffer. Diluted standards, controls, and samples were added to assay plates and incubated for 60 min at 20°C Plates were then washed, followed by addition of goat anti-human IgG-HRP conjugate (final concentration: 0.025 μg/mL) followed by another 60 min incubation. After washing, TMB substrate was added for 25 min at 20°Cin the dark. The reaction was stopped with 0.36N sulfuric acid (ELISA stop solution) (J.T. Baker; Cat# 4700–01), and absorbance was read at 450 nm and 620 nm using a SpectraMax M3 reader. Data were analyzed using SoftMax Pro 7.0.3 GxP; antibody concentrations were reported as binding antibody units (BAU)/mL. Semi-automated ELISA workflow The semi-automated workflow followed the same basic protocol with the following modifications: Plates were coated using a peristaltic-pump dispenser and stacker. After coating with 0.45 μg/mL NC antigen, plates were sealed and returned to the output stack, then restacked and incubated at room temperature for at least 4 h. Blank, non-coated plates were used on top and bottom to minimize evaporation. Post-incubation, plates were washed and blocked using a syringe pump dispenser (Pump A), then sealed and stored at 2–8°C. Standards, controls, and samples were prepared as described in conventional workflow section and stored at 2–8°C until use (between days 2 and 7). On the day of the assay, stored coated and blocked assay plates were equilibrated at 20°C for 30 min. Assay plates were washed, and pre-diluted reagents were transferred from storage to assay plates using an automated multichannel pipette (Integra VIAFLO). Sample incubation time was adjusted based on total plate count using the formula: 60 min – (1.25 min × total plates). Reagents (HRP conjugate, TMB substrate, and 0.36N sulfuric acid) were loaded into syringe-pump bottles B, C, and D, respectively. An amber glass bottle was used for TMB to protect it from light. To ensure proper priming and account for the dead volume of syringe pump bottles, 40 mL excess volume was used per reagent. After sample incubation, the semi-automated ‘ELISA’ protocol was initiated on the washer-dispenser connected to the stacker (see Supplementary Fig. 1). Blank or non-coated, dark-bottom plates were used on top and bottom to minimize evaporation and protect the assay plates from direct exposure. Immediately after the addition of 0.36N sulfuric acid (ELISA Stop Solution) to the final plate, absorbance was read at 450 nm and 620 nm. Data was analyzed using SoftMax Pro 7.0.3 GxP. ELISA Protocol Overview for Semi-Automated Method: Prime syringe pump B with goat anti-human IgG antibody-HRP conjugateDeliver plate from input stackWashDispense goat anti-human IgG antibody-HRP conjugate (100 μL/well)Return plate to output stack and restack in input stackIncubation delay: 60 min – 1.25 min × (total plates +1)Prime syringe pump C with TMB substrateWashDispense TMB substrate (100 μL/well)Return and restackIncubation delay: 25 min – 1.25 min × (total plates +1)Prime syringe pump D with ELISA stop solution.Dispense ELISA stop solution. (100 μL/well) Prime syringe pump B with goat anti-human IgG antibody-HRP conjugate Prime syringe pump B with goat anti-human IgG antibody-HRP conjugate Deliver plate from input stack Deliver plate from input stack Wash Wash Dispense goat anti-human IgG antibody-HRP conjugate (100 μL/well) Dispense goat anti-human IgG antibody-HRP conjugate (100 μL/well) Return plate to output stack and restack in input stack Return plate to output stack and restack in input stack Incubation delay: 60 min – 1.25 min × (total plates +1) Incubation delay: 60 min – 1.25 min × (total plates +1) Prime syringe pump C with TMB substrate Prime syringe pump C with TMB substrate Wash Wash Dispense TMB substrate (100 μL/well) Dispense TMB substrate (100 μL/well) Return and restack Return and restack Incubation delay: 25 min – 1.25 min × (total plates +1) Incubation delay: 25 min – 1.25 min × (total plates +1) Prime syringe pump D with ELISA stop solution. Prime syringe pump D with ELISA stop solution. Dispense ELISA stop solution. (100 μL/well) Dispense ELISA stop solution. (100 μL/well) Assigning acceptance ranges for controls (CN, CPL and CPH) The acceptance ranges for the internal controls (CN, CPL, and CPH) were established in accordance with our internal Standard Operating Procedures (SOPs), which are aligned with Clinical and Laboratory Standards Institute (CLSI) guidelines and previously described methodologies (Haynesworth et al., 2024). Briefly, a minimum of 96 replicates of each control were evaluated under controlled experimental conditions. Testing was performed by two independent analysts over three consecutive days, with 16 replicates per analyst per day, to capture both intra- and inter-operator variability. For the positive controls (CPL and CPH), control ranges were determined using a non-parametric approach. The 25th percentile (Q1) and 75th percentile (Q3) were calculated, and the interquartile range (IQR = Q3 − Q1) was used to define acceptable limits based on Tukey's method for outlier detection, as follows: Lower limit = Q1 − (1.5 × IQR)Upper limit = Q3 + (1.5 × IQR) Lower limit = Q1 − (1.5 × IQR) Lower limit = Q1 − (1.5 × IQR) Upper limit = Q3 + (1.5 × IQR) Upper limit = Q3 + (1.5 × IQR) For the negative control (CN), all replicate measurements were assessed relative to the lower limit of quantification (LLOQ). If all values were below the LLOQ, the LLOQ was assigned as the upper acceptance limit for the negative control. These predefined acceptance ranges were subsequently applied to all assay plates to ensure consistency and validity of assay performance. Statistical analysis Data organization, statistical analysis, and data presentation were conducted using Microsoft Excel and GraphPad Prism 10. Results Assay performance Each assay plate included three pre-validated internal controls: Control Negative (CN), Control Positive-Low (CPL), and Control Positive-High (CPH). Acceptable ranges were CN <24.6 BAU/mL (Lower Limit of Quantitation [LLOQ]), CPL=30.7–58.2 BAU/mL, and CPH = 158.2–585.2 BAU/mL. Any plate failing these criteria was excluded. All plates from both workflows met these criteria (Fig. 1, Supplementary Data File 1), and all sample data were included in further analysis. Assay precision and accuracy Accuracy of the semi-automated workflow was evaluated using % error relative to the conventional workflow. All 16 samples, covering the full dynamic range, showed % error < 20.0% (Fig. 2, Supplementary Data File 1). Assay precision was measured using Relative Standard Deviation (RSD) across all plates. All 16 samples showed RSD < 25.0% for both workflows (Fig. 3, Supplementary Data File 1). Discussion The semi-automated ELISA workflow described in this study substantially increased assay throughput, doubling daily plate capacity per analyst (Table 1) while reducing hands-on time from approximately 5.5 h to 1 h. These estimates, derived from direct observation under routine laboratory conditions, reflect active operator time for reagent preparation, plate washing, reagent addition, and plate handling. Despite the major assay changes that led to these operational gains, assay performance remained consistent with the conventional workflow (Fig. 3). Precision and accuracy were maintained across all samples, and although one low-response sample (sample#2) in the conventional workflow produced an outlier, overall variability for this sample remained within acceptable limits (RSD < 25%). No plate failures or control deviations were observed in either workflow. While this precluded a direct comparison of error or failure rates, the reduced hands-on intervention in the semi-automated workflow is expected to minimize operator-dependent variability. Further studies incorporating larger number of samples and other conditions would be useful to systematically evaluate differences in assay robustness and error rates. No uniform reduction in assay variance was observed across all samples, yet several low-response samples exhibited lower variability under semi-automation (Fig. 4). This suggests that minimizing manual handling may be particularly beneficial in low-signal conditions, where small procedural inconsistencies can disproportionately affect assay performance. Overall, precision remained comparable between workflows, indicating that semi-automation preserves assay reproducibility while improving operational efficiency. Although the present evaluation focused on SARS-CoV-2 Nucleocapsid IgG detection, the workflow is broadly applicable to ELISA targeting other antigens. Fully automated ELISA platforms, such as the Dynex DSX and EUROIMMUN EUROLab Workstation, provide high-throughput, walk-away operation but require substantial capital investment, proprietary consumables, and offer limited flexibility for assay modification. These constraints can limit adoption in research, academic, early-stage translational and GCLP compliant reference laboratories. The semi-automated approach presented here offers a practical alternative, balancing cost, flexibility, and performance while enabling meaningful gains in throughput. Throughput in this workflow is ultimately constrained by assay kinetics rather than robotic capacity. Although the system can technically process 30 plates per run, operation beyond 18 plates per run is not feasible due to timing conflicts between TMB substrate incubation and ELISA stop solution addition. Microplate washing and reagent dispensing require approximately 75 s per plate; processing 20 plates would extend this step to ~25 min, resulting in prolonged TMB incubation for early plates and increasing the risk of signal saturation. Similar timing-driven constraints have been reported for other semi-automated ELISA systems, where strict incubation requirements limit achievable throughput despite available robotic capacity (Rey and Wendeler, 2012; Conkright-Fincham et al., 2022; Brescia, 2023). These findings align with our observed practical upper limit of 18 plates per run (Fig. 2). This study, however, presents some limitations. First, the evaluation was restricted to a single antigen target, and generalization to other ELISA formats requires further validation. Second, the use of pre-characterized samples may not capture the full variability encountered in broader clinical or research settings. Third, precision was assessed as overall assay variability and not partitioned into intra-plate, inter-plate, or inter-day components. Fourth, system performance at maximum capacity (≥ 18 plates per run) was not experimentally verified. Finally, although no plate failures occurred, the study was not designed to quantitatively compare error rates or robustness between workflows. Future work will include empirical evaluation of full 18-plate runs to assess potential impacts on sensitivity, precision, and dynamic range, as well as exploration of workflow refinements to further increase throughput without compromising assay integrity. These efforts will support broader implementation of semi-automated ELISA platforms across diverse antigen targets and laboratory environments. Supplementary Material Supplementary data to this article can be found online at https://doi.org/10.1016/j.jim.2026.114075.