• Sign In
  1. Requests

Blank samples used to build the calibration curve

support

View Request

view list print
Blank samples used to build the calibration curve MathieuG_KI  2026-03-03 07:01
 

I am trying to set up quantitative analysis in the "small molecule" interface.
I am working with external calibration curves and normalization using surrogate deuterated standards. I am using linear fitting, and 1/x weighting.
I have the following sample types in my batch : Standard, Blank, Double Blank, Unknown.
I want to calculate LOD & LOQ based on Blank samples.
What I do not understand is that the Blank samples are used to force the calibration curve. I would like to avoid it. I am only interested in using them to calculate LOD/LOQs, but they are now overweighing the actual standards for calibration curve regression. This seems to be generating a strong accuracy bias at low concentrations. I am attaching screenshots showing this phenomenon with the successive exclusion of low-concentration standards. Irrespective of the included standards, the blank samples are always placed exactly on the calibration curve.
Can you please help me build a calibration curve without forcing it onto blank samples?
Thanks for your help!
 
 
Nick Shulman responded:  2026-03-03 07:28
Those blank samples do not affect the slope and intercept of the calibration curve.

When a concentration has not been specified for a Replicate, the point will be plotted on the graph with a y-coordinate corresponding to the observed intensity, and whatever x-coordinate would put the point along the calibration curve.
In each of your screenshots, the blanks are being drawn at exactly the same height, but their horizontal position changes because the calibration curve is different.
The blanks do not affect the calibration curve.

However, I am a little confused by your third screenshot. In that third screenshot, there are only two standards left which have not been excluded from calibration, but the calibration curve does not exactly go through those two remaining points. I suspect that the reason it looks like that might have something to do with the rounding of numbers or the particular way that Skyline is drawing that line on the graph. If there are only two external standards, and Skyline can choose any slope or intercept, I certainly would expect Skyline to choose the slope and intercept which causes the curve to go through those two points.

If you send us your Skyline document I might be able to give you a better answer.
In Skyline you can use the menu item:
File > Share
to create a .zip file containing your Skyline document and supporting files including extracted chromatograms.

Files which are less than 50MB can be attached to these support requests. You can always upload larger files here:
https://skyline.ms/files.url
-- Nick
 
MathieuG_KI responded:  2026-03-03 07:58
Thank you very much Nick. I get your point. I am now comforted to know that blank samples do not impact the calibration curve.

To reply to your concern, please note that the pictures I shared were zoomed in on the low concentrations only. There are other calibration points at higher concentrations, which I did not show.

As a complement, is there a report available to assess the quality of the external calibration, which would include relative deviations between calculated and specified concentrations of standards (bias in %)? Other metrics such as S/N or peak height would also be welcome. This would help to include/exclude points of the curve based on desired specifications.

Best,

Mathieu