ImmunoAssay Analysis Pipeline
Sandwich-ELISA plate reader data → QC'd concentrations → statistics → report. Source and documentation: GitHub.
Demo study (simulated, known truth) Two cytokines, three groups, four plates with calibrator bias, edge effects and outlier wells, all corrected and QC'd. Real data: mouse serum IL-6 after sleep fragmentation Dryad doi:10.5061/dryad.tdz08kq3p (CC0). Four plates, 110 samples; most below the LLOQ. Real data: SoftMax Pro plates across two kit lots ELISAtools example plates (MIT). Triplicate outlier QC, and a ~30% kit-lot shift corrected with a bridge control. Notebook: re-analysis of the IL-6 study 4PL vs the authors' linear calibration, censoring, strict vs sensitivity analysis, biological interpretation. Notebook: rule-based QC vs the analyst, kit-lot normalization What an analyst's manual well masking does to reported precision.