small molecules: LOD, LOQ and S/N

support
small molecules: LOD, LOQ and S/N ayson  2020-04-02
 

Hello! Good afternoon! I have run standard curve on some small molecules in an Agilent QQQ. I changed my LOD and LOQ settings from the molecule settings. My questions are the following:

  1. I am using quadratic fit and not linear because most of the molecules I ran followed this pattern. But when I got my LOD's, I am getting negative values. I believe it is because of the nature of quadratic fit. Can we have absolute values for our LOD's?

  2. I am also noticing that for some of my molecules, the LOQ's are so much lower than the LOD's. Usually, LOD's are lower than LOQ's. Can you explain why I am getting this?

settings:
20% for max LOQ bias
15% for max LOQ CV
LOD as blank + 3*SD

  1. how can we export signal to noise ratio?

Thank you!!!

 
 
Nick Shulman responded:  2020-04-02
Can you send us your Skyline document?

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.

If that .zip file is less than 50MB, you can attach it to this support request.
Otherwise, you can upload it here:
https://skyline.ms/files.url

The limit of detection involves looking at the observed peak areas of the blank samples, and figuring out where that would fall on the calibration curve. It sounds like Skyline might not be correctly taking the inverse of the parabola function.

-- Nick
 
ayson responded:  2020-04-03
Thanks, Nick! I have attached the share file data. I have also attached a word document showing you the LOD and LOQ's of the molecules attached. Thanks!

looking forward to your answer,
Marites =)
 
Nick Shulman responded:  2020-04-03
Thanks for sending those files.

When Skyline is trying to find the limit of quantification, Skyline examines each concentration level, starting with the lowest concentration level.
Skyline says that the first concentration level which satisfies the criteria of maximum %bias and %CV is the lower limit of concentration.

In your case, your lowest concentration point satisfies this criteria, but your more concentrated points probably would not. The % bias part of the criteria is a measure of how close the back-calculated concentration is to what it is supposed to be. When the regression method is quadratic, the bias is able to be particularly low, because the curve has enough degrees of freedom to pass very close to where those lowest calibration points are. The same would be true if you chose a regression weighting of something like "1/x"-- the calibration curve would necessarily be able to get very close to the lowest calibration points because those points are weighted so heavily.

I will need to ask around and see whether there is something Skyline should be doing differently in this case.
-- Nick