Quality, rigor and transparency are very important values for Olink Proteomics. Consequently, all of our assays are rigorously quality controlled and our validation data is made freely available. Below you can see the details of exactly how we validate each individual assay, and the product as a whole.
The variables described below are presented for each assay on the individual biomarker pages for Olink® Explore 1536.
Validation data overview
You can download an Excel file containing all available validation data for the assays in Olink Explore 1536:
In vitro sensitivity (calibrator) curves using recombinant antigen are determined in multiplex format, for each of the 384-plex panels within Olink Explore. Note that in some cases, no suitable antigen is available and no calibrator data is presented. Limit of detection (LOD) is defined as 3 standard deviations above background and reported in pg/mL.
Example calibrator curve from assay validation
High dose hook effect
The high dose hook effect is seen when there is an antigen excess relative to the reagent antibodies, resulting in falsely low results. In such cases, a significantly lower value may lead to erroneous interpretation of results.
Therefore, the hook threshold is determined for each analyte and reported in pg/mL and this value is presented on each individual biomarker page.
The analytical measuring range is defined by the lower limit of quantification (LLOQ) and upper limit of quantification (ULOQ) and reported in pg/mL. Quantification limits of LLOQ and ULOQ are calculated using relative error <30% and CV <30%.
Example analytical measuring range data from assay validation
The levels of protein measured in a number of commercial plasma samples are presented in sample distribution plots. Healthy subjects (n=24) and samples obtained from patients with a range of diseases (n=48). These include inflammatory, cardiovascular (n=12) , autoimmune (n=12) & neurological diseases (n=12), as well as cancer (n=12). These data provide a general idea of the NPX range to expect but cannot cover all potential levels in clinical samples.
Example sample distribution plot from assay validation
Intra-assay variation (within-run) is calculated as the mean CV for 6 individual samples, within each of 7 separate runs during the validation studies. Inter-assay variation (between-runs) is calculated as the mean CV, for the same 6 individual samples, among 7 separate runs during the validation studies.
Across all 1472 assays, the mean intra-assay and inter-assay variations observed were 8% and 11%, respectively.
Inter-site variation (between-site) was also investigated during the validation of Olink Explore at three sites (two service laboratory sites and Olink R&D), to estimate the expected variations in values between different laboratories running the same samples, with different operators using different equipment. Each site was trained and instructed to perform the analysis according to the same routine. Each site performed 6-7 independent runs.
The analysis of results from these multi-site studies are currently being finalized and will be updated here as soon as they are available.
PEA and specificity – a major problem solved
The unique features of PEA technology overcome a long-standing and well recognized issue with immunoassays. With standard assays such as ELISAs, even moderate levels of multiplexing result in cross-reactivity of antibody binding and a loss of specificity of the signals that are detected (Figure A below). By virtue of PEA’s requirement for dual antibody recognition of the target protein and high-fidelity DNA hybridization and detection, any unspecific antibody binding event that may occur will not result in a readout signal (Figure B below).
Specificity validation in Olink Explore
Rigorous specificity testing was used during the development of Olink Explore 1536, both as part of the assay selection process, and afterwards as part of the formal product validation procedure.
All assays have gone through a predefined protocol of at least 3 levels of specificity testing:
First, a screen against several pools of antigens (Ag) is performed (n>100)
After removal of poorly performing antibodies, a second screen is performed using an expanded set of Ag pools
Validation of final product design against pools of carefully selected proteins (n=96) including,
15 well-known biomarkers from each of the four 384 panels
36 proteins with high homology within their protein families
In addition, all Explore assays with equivalent assays in Olink Target 96 panels (n=1118) have previously been tested for specificity against the proteins in the corresponding Target 96 panel.
Overview of specificity validation results
In total, 99.8% of assays (1469/1472) in Olink Explore 1536 exhibited no cross-reactivityaccording to the tests described. FOLR3 showed a non-specific signal against the related FOLR2 protein, CCL3 against the related CCL4 protein, and LHB against CGB. The signal contribution at endogenous levels were further investigated and more details are noted on the specific biomarker assay pages concerned.
The example graph above shows the specificity testing approach. Each assay is exposed to samples containing either subsets or all of the selected antigens. A given assay should only generate a signal when the corresponding antigen is included in the sample. The lack of unspecific signal indicates that each assay is specific for its target antigen, demonstrating the outstanding specificity of the PEA method at high multiplex degree
If you have any questions or comments about how we validate our assays, or if there is anything else we can help you with, please contact us using the link below: