LABREADOR: A zero-footprint, client-side web application for automated quality control and nonlinear regression in microplate bioassays

Marek Helczman
Journal of Immunological Methods · Elsevier · DOI: 10.2139/ssrn.6487438

Abstract

Quantitative analysis of microplate bioassays, such as ELISA, requires robust nonlinear regression tools. These are typically available either as expensive commercial software or as simple web-based alternatives with insufficient analytical capabilities. In this work, we present Labreador—a freely available web platform built on a client-side architecture that eliminates the need to transfer data to external servers. The system integrates automated selection from five regression models (including 4PL and 5PL) with the Levenberg-Marquardt optimization algorithm, weighted regression for heteroscedasticity correction, interactive outlier management, and a three-level quality control system with automatic blocking of interpolation when the calibration curve quality is insufficient (R² < 0.95). The platform&apos;s accuracy was verified on a synthetic reference dataset. Following the application of 1/Y² weighted regression and interactive outlier management, the algorithm recovered an excellent fit (R² ≥ 0.98) and the true EC₅₀ reference value of 150.0 fell within the recovered 95% confidence interval, confirming the statistical validity of the parameter estimate in the presence of 10% CV biological noise. Labreador provides computational accuracy comparable to that of commercial platforms in a free, privacy-oriented environment accessible without any technical barriers.

Read at publisher →

Similar articles