Dear Skyline Users,

We had some good questions during our sixth Skyline Tutorial Webinar so we have provided the answers!

Q: Can you describe how you normalize data in Skyline?

Ans:1. Ratio to Global Standards and 2. Ratio to heavy (or any other internal standard isotope label type). It is important to note that for targeted proteomics is it not a valid assumption that you can normalize to the median intensity peptide, as you might for a large survey/discovery experiment, where you expect the majority of peptides/proteins not to be changing. The first option, Ratio to Global Standards allows you to normalize to the sum of peak areas of a prespecified set of standard peptides, which you should verify work well (i.e. provide stable area measurements) ahead of using them in your experiment. In the data set shown in the webinar, it turned out that 2 out of 3 of the intended global standards were very early eluting and experienced serious signal loss that was not typical of the rest of the targets. The second option is much simpler, but requires a stable isotope labeled internal standard for every target you intend to measure.

Q: How do you use MSstats to compare 3 biological replicates of "Control" samples to 3 biological replicates of "Treatment01", 3 biological replicates of "Treatment02", 3 biological replicates of "Treatment03" and so on? I managed to compare control-treated but not control to different treatment conditions.

Ans: You will need to use the MSstats R interface to do this, and consult its documentation or support board (which you can find on the MSstats details page in the Skyline External Tool Store - ). The MSstats external tool Group Comparison form that can be shown from the Skyline Tools menu currently only supports pairwise comparison, the same as the new Skyline integrated group comparison support shown in the webinar.

Q: How does the standard work? I notice that ’S’ had a positive fold-change.

Ans: Great question. I believe this was due to a lack of randomization in the experiment, with all diseased samples being measured before all healthy samples in each of three blocked cycles, and the fact that 2 out of 3 of the peptides in S experienced 20- to 200-fold signal loss over the period of the experiment. Because of this, those 2 peptides were not actually used for normalization, and they are what ends up causing the ‘S’ protein to appear to have higher expression in diseased than in healthy. This is all explained in the new Group Comparison Data Processing tutorial that is now a companion to the webinar.

Q: Could you hide some of the replicates during work instead of deleting them ? So that they won't be seen but the data will be still in the document.

Ans: This has been requested before, and would clearly be helpful to people working with large data sets, say over 100 runs. We will get to this eventually. It just hasn’t made it to the top of our priority list yet.

Q: Where can we get access to the full tutorial?

Ans: There is now a link to the full tutorial on the webinar page itself, on the Skyline main page and on the tutorials page, and now here (https://skyline.gs.washington.edu/labkey/turial_grouped.url).

Q: Would there be an option to view the manually edited peaks? (Eg. the ones that are truncated , in order to mark the samples that should be rerun)

Ans: This was shown during the webinar. So, it should be possible to review the video to see how this was done using the Document Grid. It is also explained pretty thoroughly in the written tutorial.

Q: Would the statistics shown work on MS1?

Ans: Yes. Some of our first testers of this feature used it for MS1 with known expected outcomes, and they reported that it worked well for them.

Q: Could you please comment how Skyline assembles proteins from peptides in group comparison? Is it a sum of intensities of the most intense transitions from all peptides from the protein?

Ans: Close. It is the sum of all targeted transitions in the protein. We are aware that this does not always work well for MS1 where risk of interference is high enough that using the lower intensity transitions can introduce considerable error, but frequently with SRM and PRM, where the probability of interference is much lower, then using the other ions produces more precise measurements. In the future, we will likely add an option to use the top N transitions, with separate values for MS1 and fragment ions.

Q: How does Skyline calculate the values for truncated peaks? What's the algorithm behind?

Ans: The definition of a “truncated peak” in Skyline is one where one of the boundaries is at either extreme edge of the chromatogram, and the intensity at that extreme edge is greater than 1% of the total peak height higher than the intensity where the chromatogram crosses the interior boundary. This is pretty conservative, and it should be possible to find cases where you think Skyline is calling a peak truncated, and therefore excluding it from quantification where you feel the peak could be used. You can, however, easily correct this by adjusting the integration boundary at the edge of the chromatogram over by one time interval, so that it is no longer at the extreme edge. If it really is not truncated, this should not matter much. We decided to err on the side of caution, rather than letting peaks missing significant area due to truncation impact quantification.

Q: And how does it calculate the area value for truncated peak?

Ans: Skyline calculates the area for a truncated peak the same way it calculates area for any peak, which is described in a Tip on the Skyline web site here:  https://. Briefly, it uses the trapezoidal area under the chromatogram, minus the background, estimated as the rectangular area between the integration boundaries, the x-axis and a horizontal line from the lowest point where the chromatogram crosses an integration boundary. For a truncated peak, this last boundary is by definition defined by the intensity of the chromatogram where it crosses the interior boundary, and not the boundary at the extreme edge.

Q: Can we modify retention time window to include truncated peak for individual file?

Ans: Well, in the case of SRM (as with the data shown) or PRM, you cannot, by definition, because the chromatograms represent the data that was collected by the instrument during schedule acquisition, and there is nothing else to be shown. In theory, with MS1 or DIA data you should be able to extend the chromatograms to include a wider time range. However, Skyline currently only supports making this modification by modifying the document settings and then re-importing the data. I hope we will be able to make this more convenient in the future.

Q: How can we set ppm (vs. m/z) mass tolerances for peptide library and transition matching?

Ans: At present, Skyline only supports using m/z for spectral library mass tolerance in transition matching (see Transition Settings – Library tab). It can be set to very small values that work nearly as well as PPM values, but which are constant for all m/z values, unlike PPM which will vary with the m/z value being matched. This is a good request and hopefully it can be done fairly soon by adding a “High accuracy” checkbox beneath the field in the Transition Settings – Library tab that toggles between PPM and m/z. Thanks for your feedback.