Blog
Building a Calibration Curve for Peptide Quantification: Standards, Internal Standards, Weighting and LLOQ
Anyone who has measured how much peptide is in a sample has built a calibration curve, and many have built one badly. A straight line with a high correlation coefficient looks reassuring, but it can hide a poor method. A few design choices make the difference between a curve that looks good and one that is right.
Key takeaways
- Make standards from an independent weighing, in the same matrix as your samples.
- Use at least six non-zero levels spanning the range you will measure.
- Add an internal standard, ideally a stable-isotope-labeled version of the peptide.
- Use weighted regression, such as 1/x or 1/x squared, because errors are larger at high concentrations.
- Judge the curve by back-calculated accuracy, not just R squared.
Start with the standards
A calibration curve is only as good as the standards. Prepare a stock from carefully weighed material of known purity and net peptide content, as described in our salt forms post and our A280 guide. Make serial dilutions into the same matrix as the samples, for example buffer or plasma, because matrix changes the signal. Use low-binding tubes, as in our adsorption post. Include a blank with no peptide, and a blank with only the internal standard.
The internal standard
An internal standard is a compound added at the same amount to every standard and sample. The result is expressed as a ratio of the peptide signal to the internal standard signal, which cancels out variations in injection volume, losses and instrument drift. The best choice is a copy of the peptide labeled with heavy carbon or nitrogen, which behaves almost identically in sample preparation and chromatography but is separated by mass. A close structural analog is the next best option.
Weighted regression
Measurement error in LC-MS usually grows with concentration, so the high points scatter more in absolute terms. An ordinary least squares line lets those high points dominate, which makes low concentrations inaccurate. Weighting each point by 1 over x or 1 over x squared gives the low end its proper influence. Use the weighting that gives the best back-calculated accuracy across the range and justify it.
Judging the curve
- Back-calculate each standard from the fitted line.
- Common acceptance limits: within 15 percent of the nominal value, and within 20 percent at the lowest level.
- Check that at least three quarters of the standards pass.
- Do not drop points without a documented reason.
- Check for carryover by running a blank after the highest standard.
LLOQ and detection limits
The lower limit of quantification, or LLOQ, is the lowest concentration measured with acceptable accuracy and precision, typically within 20 percent. A related rule of thumb from validation guidelines sets the limit of detection at about 3.3 times the noise divided by the slope, and the limit of quantification at about 10 times that ratio. Quantification limits should be demonstrated with real samples, not just calculated.
Where it fits
Calibration curves underpin stability work such as serum stability assays, binding studies and purity quantification. Method choice, from detection wavelength to mass spectrometry, should come first, with the curve built around it.
The bottom line
A good calibration curve is the foundation of every number in a peptide study. Design it with independent standards, an internal standard, the right weighting and honest acceptance criteria. For the documentation we provide on our lots, see the lab reports page.
Related reading
- Serum and plasma stability assays
- Concentration from A280
- HPLC versus mass spectrometry
- Browse the full Research Library
Explore Lucid Aminos
- Shop all research compounds
- Read the lab reports (COA)
- Monthly Research Panel
- Frequently asked questions
- Partner program
- About Lucid Aminos
- Contact us
For laboratory research use only. Not for human or veterinary use, and not intended to diagnose, treat, cure or prevent any condition. This article is educational and is not medical advice.