elisa-assay-design

Overview

elisa-assay-design covers developing enzyme-linked immunosorbent assays from scratch or optimizing kits: format selection, antibody pairing, blocking and buffer optimization, standard curves, and the validation parameters (accuracy, precision, LLOQ/ULOQ) that make an ELISA a measurement tool rather than a color-change ritual.

When to use

• Choosing ELISA format: direct, indirect, sandwich, competitive.
• Antibody pair selection and validation for sandwich ELISAs.
• Optimization: coating, blocking, detection antibody, substrate, buffers.
• Standard curves: 4PL/5PL fitting, weighting, acceptance criteria.
• Validation: precision, accuracy, dilutional linearity, spike-recovery, LLOQ/ULOQ.
• Troubleshooting: high background, low signal, hook effect, edge effects.
• Transferring assays between labs or to kit format.

Core concepts

• Format selection. Sandwich (capture + detection antibodies — most specific and sensitive, needs two non-competing antibodies); indirect (antigen-coated, labeled secondary — simpler, less specific); competitive (labeled vs unlabeled antigen compete — for small molecules that can't be sandwiched, e.g. steroids, drugs); direct (labeled primary — fast, insensitive). Small analytes force competitive; everything else prefers sandwich.
• Antibody pairs. The critical reagent decision: capture and detection antibodies must bind non-overlapping epitopes (test empirically — datasheet claims aren't enough); validate specificity against related analytes (cross-reactivity panels); monoclonal-monoclonal for consistency, with polyclonal detection as sensitivity fallback. Lot-to-lot variation is real — qualify new lots against old.
• Blocking and buffers. Block unbound surface (BSA, casein, non-fat milk — test which; milk contains biotin, which matters for streptavidin systems); sample diluent should mimic matrix; wash thoroughly (insufficient washing is the top cause of high background). Heterophilic antibody interference (human anti-mouse antibodies) needs blockers in clinical samples.
• Standard curves. Recombinant/known-concentration standards in matrix-matched diluent; 4PL or 5PL fit (5PL for asymmetric curves); 1/y² weighting (variance scales with signal); 7-8 points spanning the range plus blank. Curve acceptance: back-calculated standards within ±20% (±25% at LLOQ/ULOQ), R² not sufficient alone — inspect residuals.
• Validation parameters. Precision (intra/inter-assay CV: ≤20%, ≤25% at limits); accuracy (spike-recovery 80-120% in matrix); dilutional linearity (serial dilutions of high samples parallel the curve); LLOQ/ULOQ (lowest/highest standard meeting precision+accuracy); selectivity (related analytes, hemolyzed/lipemic samples); stability (freeze-thaw, bench-top).
• Hook (prozone) effect. Very high analyte saturates both antibodies → falsely low signal. Suspect when strong positives read low; test by diluting samples. Sandwich assays for analytes with huge dynamic ranges need dilution protocols built in.
• Matrix effects. Serum/plasma components shift the curve vs buffer standards. Mitigate: matrix-matched standards, minimum required dilution, spike-recovery validation in the actual matrix. An assay validated in buffer and run in serum is unvalidated.
• Controls. High/medium/low QCs on every plate, tracked on Levey-Jennings charts; plate acceptance by QC rules (not by whether the samples "look right"). Trending QCs catch reagent degradation before it ruins data.

Practical workflow

1. Format and reagents. Choose format by analyte size; screen antibody pairs empirically; qualify lots.
2. Checkerboard optimization. Titrate capture vs detection antibody (and sample vs conjugate) in a matrix; select for signal:background ratio, not just signal.
3. Buffer optimization. Blocking agent, wash stringency, incubation times/temperatures — one factor at a time after the checkerboard.
4. Standard curve. Matrix-matched; 4PL/5PL with weighting; define acceptance criteria.
5. Validate. Precision, accuracy, linearity, LLOQ/ULOQ, selectivity, stability — the full panel before generating study data.
6. Run with QCs. Every plate: standards + QCs + blanks; Levey-Jennings tracking; documented plate-acceptance rules.
7. Troubleshoot systematically. High background → washing/blocking/antibody concentration. Low signal → antibody activity, substrate, incubation. Edge effects → plate sealing, temperature uniformity, avoid outer wells for critical samples.

Common pitfalls

• Antibody pairs assumed (not tested) to be non-competing.
• Standards in buffer, samples in serum (matrix mismatch).
• Hook effect unrecognized (high samples reading falsely low).
• No dilutional linearity testing.
• QC trending ignored until assays fail outright.
• Lot changes without bridging qualification.
• Curve fit by R² alone without back-calculation checks.