| Hi Brendan,
I was working in Skyline yesterday and I notice something a bit strange. In the PowerPoint file I’m sending to you, you can see the issue I’m having. I created a Peak Area Report so that I can export things out of Skyline to work with. For this particular peptide, I have an isotopically labeled peptide standard that I spiked in with the native peptide. I’ll give you this slight caveat early. The samples that I’m working with for this project, don’t give the cleanest transitions.
As you can see in the first slide, the light peptide is clearly in the sample at a higher amount than the heavy peptide. However, when I export the numbers out using my custom report, I get a L/H Ratio that tells me just the opposite. I checked this by manually integrating the peaks in Xcalibur and got a L/H Ratio that was much closer to what I see visually. I know the reason for this is that Skyline probably is only choosing part of the light peak to integrate or not choosing the right peak at all, but my problem is that I can’t see where Skyline is choosing to integrate. Is there a way to see this in Skyline? Thanks Brendan! |
|
| |
| |
| maccoss responded: |
2010-07-22 10:54 |
It looks like there is a high background. So if the background is subtracted at the valleys from the two peaks adjacent to the main red peak it will reduce the area of the red peak substantially. Although I'm not sure if this is the case or if it is selecting the integration bounds incorrectly. Where are the integration boundaries placed? I don't see them on the current plot.
Cheers,
Mike |
| |
| Brendan responded: |
2010-07-22 11:06 |
Hi Rudy,
Skyline should be showing you the integration boundaries it has picked, and if you are using v0.6, then it should be picking the same integration boundaries for both light and heavy precursors, unless you explicitly set your isotope labeling modification to another option (e.g Before, Overlapping, Unknown).
If you are not seeing the integration boundaries, try right-clicking on the chart, and making sure that the Peak Boundaries menu item has a check mark beside it. If it doesn't, then you must have turned this off at some point. Just click the menu item and the boundaries should reappear.
My guess is that the boundaries will appear for both peaks between the two humps on either side of the red peak. This would cause the background for the blue peak to be set near zero and the background for the red peak to be up around 200 on the y-axis. Definitely have a look at the TotalBackground values for the two precursors. Given what you explain, I would expect the red peak to have a much higher background area.
Skyline calculates the background as the rectangular region between the peak boundaries, the x-axis, and a horizontal line crossing the lower of the two intersections between the peak boundaries and the chromatogram.
If none of this turns out to be the case, then we may have to arrange for me to look at the file to understand what is going on.
Let me know what you see.
--Brendan |
| |
| rjalvarado responded: |
2010-07-22 11:08 |
| Expires: 2010-07-22 |
Hi Mike,
Thanks for the response. I didn't know that there was a background subtraction algorithm being applied? I also didn't know how to show the integration boundaries until I saw your response. I enabled them and it looks like it is integrating the correct peak, so maybe the background subtraction is the answer to my problems. Is there any way to turn this off or is it something that is static? Thanks!
~Rudy |
|
| |
| Brendan responded: |
2010-07-22 11:30 |
Hi Rudy,
Hi Rudy,
You can't turn off background subtraction, and you can't correct where Skyline sets the background level separately from the peak boundaries. The background level is always taken as the lower of the two intersections between the peak boundaries and the chromatogram. In this case, the only way to change the background level would be to include at least one of the smaller humps on either side of the red peak. I would choose the one on the right, since the blue peak also has some tailing to the right, and extending the right side boundary to where both peaks go nearly to zero is likely to have the smallest impact on your overall ratio. At least it should cause the red peak to start showing up as having more area than the blue one.
You can just click on the right boundary and drag it to where you want it, though you might have to do this for both peaks. Or you can click and drag below the x-axis to define new boundaries for both peaks.
Give it a shot, and let us know how it works.
--Brendan
P.S. - Sure wish I knew how you ended up with the peak boundaries hidden without your knowing about the menu item to do that. I checked the default settings, and they do specify showing as the default. |
| |
| maccoss responded: |
2010-07-22 11:58 |
The peaks look very different in the unsmoothed data than the smoothed data (original chromatograms you showed). I don't think you'd want to eliminate the background subtraction. The background subtraction is meant to partially account for background interference but as you get to low abundance species the interference is going to contribute more and more to the peak area. With the boundaries set like they are in Figure 3, what is the ratio now?
-Mike |
| |
| Brendan responded: |
2010-07-22 12:16 |
Hi Rudy,
I guess I didn't notice that you had posted a new screen shot with your last post. It doesn't look like the peak you are showing is the one Skyline is integrating, otherwise it would have black dashed borders (not gray) and an arrow-head pointing at the apex of the peak.
You might also want to right-click on your chromatogram graph, and choose Retention Times and then click All also, to get the retention time labels above peaks on your graphs. Still not sure how your chromatogram graphs ended up so stripped of the usual charting options without your knowing.
Then press F11 (View / Auto-Zoom / Best Peak on the menu), and Skyline should zoom into the peaks it has chosen to integrate. If you disagree, and you really want the peak showing in slide 3, you can zoom/pan back to this peak and click and drag below the x-axis to set the boundaries to what you think they should be.
We are working on a tutorial covering integration that should be ready sometime this summer.
--Brendan |
| |
| rjalvarado responded: |
2010-07-22 12:22 |
Hi Mike,
Those were the original boundaries set in the PowerPoint file that I sent to you earlier. I was just able to turn them on this the second file, so the ratios are the same. The peaks do look drastically different when smoothing is and isn't applied, which is one reason why I turned on the smoothing. The other is to decrease the amount of bias that I introduce. Is there any way for us to input our own peak smoothing algorithm or are you planning to add others options in the next release?
~Rudy |
| |
| rjalvarado responded: |
2010-07-22 12:24 |
Hi Brendan,
You were right. The peak that Skyline chose to integrate wasn't the correct peak. I dragged the integration boundaries over my peak, but some reason, it still isn't integrating the light peak correctly. |
|
| |
| rjalvarado responded: |
2010-07-22 12:31 |
| By the way, I just wanted to tell you guys how AWESOME I think Skyline is! I've already taught a couple SRM classes where I have fully endorsed Skyline and used it in the class. It is very user friendly and you guys offer great support for it. Keep up the good work! |
| |
| Brendan responded: |
2010-07-22 12:42 |
Hi Rudy,
Thanks for your support. Slide 4 is starting to look like what I would expect to see, except that the red peak still isn't showing black integration boundaries, meaning its integrated peak is still somewhere else. Is this what you got clicking and dragging beneath the x-axis? You should also be able to click on the red 53.3 label, now that it is showing, but it doesn't look like it will be using boundaries that match the blue peak at the moment, and you really want them to match. The easiest way to get that is usually clicking and dragging beneath the x-axis, but if that isn't working you can also click on the dashed lines and drag them.
Maybe we should arrange a WebEx session for you to show me the issues you are having. Sounds like it may turn up a bug that I would want to fix.
Thanks for your detailed feedback. It has taken a lot of great feedback from MS instrument operators to get Skyline to where it is.
--Brendan |
| |
| rjalvarado responded: |
2010-07-22 12:48 |
Hi Brendan,
The peak boundary is there, it is just obstructed from the other boundary on that same side. In slide 4, the red arrows show the integration boundaries and the orange arrow shows the other boundary that is hiding the integration boundary. Does that make sense?
I've also included another slide in which both integration boundaries are visible. I'm hoping that was what you were instructing me to do right?
~Rudy |
|
| |
| Brendan responded: |
2010-07-22 12:59 |
Hi Rudy,
In slides 3 and 4, the gray boundary showing that is the same height at the red peak is not being used in the integration of the peak for the red chromatogram, because it is not black, and the red peak does not have a black arrowhead pointing at it, as the blue peak does.
In slide 6, however, this issue seems to have been corrected. In slide 6, both peaks have black dashed boundaries, and a black arrowhead. These are the peaks Skyline is integrating. But, it does look like you have set different boundaries for the two peaks, when it would be better to have them be the same, as you were showing in slide for with the left-side black and gray boundaries.
All gray dashed lines are simply showing you where Skyline would integrate, if you click on the label surrounded by those boundaries (i.e. the red 53.3 in slide 4).
Hope this helps. I am feeling like we really need that tutorial...
--Brendan |
| |
|
|
|
|
|
Peak Area Issue.pptx