Skip to main content

Table 2 Parameters for partial assay validation

From: Development of a novel, sensitive translational immunoassay to detect plasma glial fibrillary acidic protein (GFAP) after murine traumatic brain injury

Parameter Definition Experimental details Sample type Technical replicates Deviation from Andreasson et al
Robustness The ability of the assay to accurately and precisely measure the analyte under small protocol variations [23]. Modify the incubation duration of capture antibody (1 h vs. 2 h vs. overnight), sample (2 h vs. 1 h vs. overnight), and detection antibody (1 h vs. 30 min) parallel assays. Analyze %CV and %recovery. Plasma pools (low GFAP: 3–6-month-old wildtype mice, intermediate GFAP: 9–24-month-old APP/PS1 mice and 18–24-month-old wildtype littermates, high GFAP: 6 h post-TBI). 2 No deviations.
Precision The variation in analyte measurement within a single assay (repeatability, intra-assay variation) and between independent assays performed on different days (intermediate precision, inter-assay variation) [23]. Repeat the assay with the same specimens on 5 independent days. Analyze intra-assay and inter-assay %CV. Plasma pools (low GFAP: 3–6-month-old wildtype mice, intermediate GFAP: 9–24-month-old APP/PS1 mice and 18–24-month-old wildtype littermates, high GFAP: 6 h post-TBI). 5 No deviations.
Limits of quantification The lowest and highest analyte concentrations in plasma that can be measured by the assay with a given level of precision (< 20% CV) [23]. LLOD: Assay replicates of diluent, calculate LLOD as 2.5 SD above the blank. Diluent (1% BSA in PBS). 16 No recommendation available in Andreasson et al [23], followed Meso Scale Discovery recommendations.
LLOQ: Assay replicates of diluent, calculate LLOQ as 10 SD above the blank. Diluent (1% BSA in PBS) 16 No deviations.
LLOQ: Assay specimens with very low GFAP concentrations. LLOQ = lowest measured concentration with duplicates < 20%CV. Plasma specimens with low expected concentrations (n = 8, 3–6-month-old wildtype mice). 2 No deviations.
ULOQ: Assay specimens with high GFAP concentrations. ULOQ = highest measured concentration with duplicates < 20%CV. Plasma specimens with high expected concentrations (n = 8, 6 h post-TBI). 2 No deviations.
Dilution linearity The ability of the assay to accurately and reliably detect the analyte in plasma spiked with the calibrator at a very high concentration after dilution [23].
The ability of endogenous analyte in plasma to be detected at various dilutions accurately and reliably [28].
Spike plasma specimens with recombinant protein 120-fold above the estimated ULOQ (2,000,000 pg/mL), perform serial dilutions until theoretical concentration is below LLOQ (1-fold to 1,000,000-fold). Analyze %CV and %recovery. Plasma specimens with low expected concentrations (n = 3, 3–6-month-old wildtype mice). 2 No deviations.
Parallelism Comparison of the signal vs. dilution factor response of the calibrator and endogenous analyte in plasma [23]. Perform serial dilutions (1-fold to 64-fold) of specimens with high expected concentration. Analyze %CV and %recovery. Plasma specimens with high expected concentrations (n = 3, 6 h post-TBI). 2 3 specimens were used rather than the recommended 4 specimens, no other deviations.
Perform serial dilutions (1-fold to 64-fold) of specimens with high expected concentration, analyze response of signal to dilution factor compared to response of recombinant protein. Definition from Sweeney et al [28]. 3 specimens were used rather than the recommended 5 specimens, no other deviations.
Recovery Ability to accurately and reliably measure the concentration of plasma spiked with calibrator [23]. Divide specimens into 4 aliquots, spike 3 of the aliquots with recombinant protein at concentrations across the range of the standard curve (100, 1000, and 10,000 pg/mL), add an equivalent volume of diluent to the fourth aliquot. Analyze %CV and %recovery. Plasma specimens with low (n = 2, 3–6-month-old wildtype mice) and high (n = 3, 6 h post-TBI) concentrations. 1 No deviations.
Stability The stability of the analyte in plasma after a given number of freeze-thaw cycles or a given amount of time at a given temperature [23]. Divide specimens into 19 aliquots. Aliquot 1 store at − 80 °C. Aliquots 2–6 exposed to 1, 2, 3, 5, or 7 freeze-thaw cycles (2 h at RT then store at − 80 °C), respectively. Aliquots 7–12 store at RT for 1, 2, 4, 24, 72, and 168 h, respectively. Aliquots 13–18 store at 4 °C as above. Aliquot 19 store at −20 °C for 1 month. Assay together and analyze %CV and %recovery. Plasma pools (low: 3–6 month- old wildtype mice, intermediate: 9–24-month-old APP/PS1 mice and 18–24-month-old wildtype littermates, high: 6 h post-TBI). 2 No deviations.
Selectivity The ability to measure the analyte in the presence of other substances expected to be present [23]. A) Compare assay performance in plasma and serum.
B) Measure the analyte in specimens with hemolysis.
A) Plasma and serum from sham (n = 6) or TBI (n = 7) mice collected 6 h post-TBI.
B) 5%, 25%, or 50% red blood cells spiked into plasma from sham (n = 3) or TBI (n = 3) mice collected 6 h post-TBI. Spiked specimens were frozen at − 80 °C then thawed and assayed.
2 Andreasson et al recommends spiking in substances with similar physiochemical structure.
Pre-analytical factors The stability of the analyte with respect to blood collection and processing methods. Blood was collected by saphenous vein and cardiac puncture. Blood from cardiac puncture was divided into 2 aliquots. One aliquot was centrifuged to plasma < 1 h after collection and one aliquot was incubated on ice 4 h before centrifugation. Plasma from sham (n = 3) or TBI (n = 3) mice collected 6 h post-TBI. 2 No recommendation available in Andreasson et al [23].
  1. LLOD lower limit of detection, LLOQ lower limit of quantification, ULOQ upper limit of quantification, TBI traumatic brain injury, CV coefficient of variation, SD standard deviation, RT room temperature