Detector Linearity in Peptide Analysis
Linearity evaluates whether analytical response changes predictably with analyte concentration across a defined range.
Calibration model
A calibration curve relates concentration to response. Linear regression is common, but the model must fit the data and intended range.
Correlation is not enough
A high correlation coefficient can occur even when:
- low-end bias is large
- variance increases with concentration
- one high point dominates the fit
- residuals show curvature
- back-calculated values are unacceptable
Weighting
Weighted regression can improve fit when variance is not constant across the range. The weighting model should be justified.
Range
The validated range is not simply the lowest and highest calibration points. Accuracy, precision, and response behavior must be acceptable throughout.
Frequently asked questions
Is R² = 0.999 automatically acceptable?
No. Residuals and back-calculated values also matter.
Can detector saturation occur?
Yes. High concentration can flatten response.
Should impurity methods be linear near the specification?
Yes. Performance should be demonstrated near reporting and specification levels.
Can one calibration cover every application?
No. Range and model must match intended use.
Key takeaways
Linearity is a property of the complete method and range. Correlation alone does not establish quantitative reliability.
References
- ICH Q2(R2). Validation of Analytical Procedures.
- ICH Q14. Analytical Procedure Development.
- FDA. Analytical Procedures and Methods Validation for Drugs and Biologics.
TSMS Labs educational disclaimer: For laboratory research and educational purposes only. Not for human consumption. This content is not medical, clinical, or regulatory advice.