Hi Jake,
Yes. Skyline does background subtraction. So the reported peak areas are always the raw integrated area minus the background area. The background subtraction is extremely important in some case and less important in others.

A reasonably comprehensive explanation of how Skyline calculates these values can be found in this support thread:

https://skyline.gs.washington.edu/labkey/announcements/home/support/thread.view?entityId=4b24cd1b-ab9a-102e-87a2-4c1490ad0666&_docid=thread%3A4b24cd1b-ab9a-102e-87a2-4c1490ad0666

Some day I will get around to writing this up more formally.

--Brendan

Good afternoon Brendan,

  Thanks for the information about background subtraction in Skyline. For our study, it seems that the raw peak areas would be more applicable, since we define the LOQ by the mean and standard deviation of the signal in triplicate blank runs. These results are very sensitive to changes in the integration window due to the resulting changes in the background area. Does Skyline have functionality for this type of LOQ calculation?

Thanks,
-Jake

Hi Jake,
You should expect CVs on noise not to be that good. AB SCIEX is pretty aggressive about attempting to zero out noise, though. So, you may be just fine without background subtraction. Up to you.

At present Skyline does not have its own calibration curve or LOQ calculations, though we expect to rectify that in 2013, and we are currently working with the Carr lab on integrating some of their work in this area.

It is always possible to export the raw numbers in a report, and pretty easy to sum Area and BackgroundArea, if you feel that gives you better results in your calculations.

Hope that helps. Good luck with your experiments.

--Brendan

Good morning Brendan,

  How can I export the raw peak areas in a report? For most of the results, I can calculate the raw peak area by adding the reported peak area to the background area. But I'm assuming this won't work for those results with a reported peak area of 0, since the background area is larger than the raw peak area.

Thanks,
-Jake

Hi Jake,
That might just be a bug I can fix relatively easily. I was just working in the peak integration code this week and noticed an odd comment about our peak integration library sometimes producing negative areas, followed by code that ensures the minimum area value is zero. Sounds like this is causing us to lose some information.

My suspicion is that background is always calculated as a rectangular area, since that is how it is defined in its most basic form (see previously reference support thread). In the majority of cases that is just fine, but your post makes me realize that my expectation of what the integration library is doing and what is actually doing in more degenerate cases may be totally different.

I have attached a little doodle I created in paint to illustrate the problem.

- What I had been expecting was that Background would be the violet shaded area, and Area would be the pink shaded area.
- What I now think is actually happening is that Background is the entire area of the black rectangle, and Area is the pink shaded area minus the gray shaded area.

Obviously the latter allows both the Area to be negative, and the background to be larger than the integrated area between the integration boundaries.

It sounds like it might be necessary to make the integration calculation a little more complicated to make it return the values as I was originally expecting until you pointed this out.

Of course, this only makes a difference when an important amount of area is below the lower of the two intensities where the chromatogram crosses the integration boundaries.

Thanks for digging into this one.

--Brendan