Guides · BioStudioAI

Why ELISA statistics usually belong on the log scale

Last reviewed Author: BioStudioAI

Answer first. Concentration errors in immunoassays are proportional, so log transformation stabilizes variance and makes fold differences symmetric. Test on log concentrations and report geometric summaries when appropriate.

Concentration errors in immunoassays are proportional, so log transformation stabilizes variance and makes fold differences symmetric. Test on log concentrations and report geometric summaries when appropriate.

Decision table

SituationAction
Controls pass, in rangeReport with the pre-set analysis
Controls pass, out of rangeDilute/rerun into range
Controls failFix the named step; repeat

Mistakes that change the answer

Choosing criteria after results, testing technical replicates as biological units, and keeping only summaries. Use the Tools calculators for arithmetic and the Protocols for bench work. BioStudioAI automates this analysis with QC and an audit trail: /platform/.

References

  1. Assay Guidance Manual · NCBI Bookshelf (2024)

Do this analysis in BioStudioAI

Upload the raw file and BioStudioAI fits, quality-checks and reports it with an audit trail.

See the platform →

Related

4PL vs 5PL vs linear: choosing a standard-curve model

Choose the simplest curve model whose residuals are patternless and whose back-calculated standards recover within your limits. 4PL handles most plate

Back-calculating concentrations and dilution factors

Back-calculate each sample from the fitted curve, multiply by its dilution factor, and accept only values that fell inside the curve's validated range

ELISA Data Analysis: The 4PL Standard Curve, Step by Step

How to analyse ELISA data with a four-parameter logistic (4PL) fit: blanks, curve quality, dilutions, CV limits, and concentrations you can defend.

ELISA