Thanks for your post. We did an experiment like this back in 2010 (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2813958/). This method gave some interesting results in terms of the number of peptides that could be qualitatively detected in a run. I believe that Skyline should be able to analyze data collected in this way but it would require you to come up with how you are going to shift the m/z range over time and acquire the data.

That said we haven't done much with this in recent years because we felt an advantage of DIA over PRM was the simplicity and generalization of a data acquisition method to most samples. Once we get into scheduling the RT we have to come up with strategies to shift the location of the isolation windows pending any RT shift between runs. If we don't then we face potentially losing peptides in some runs but not others because they fall in and out of the isolation range at a given time.

Our priorities have been less about increasing the number of peptides that can be detected in a run and more about improving the quantitative capabilities across many runs. It would be great if shifting the location of the isolation windows improved both.

Let us know how your experiments work.
Mike

Skyline can definitely handle this kind of data. You would have to generate the method yourself, but once you have acquired data with a method like this, you would just have to make sure you use a DIA isolation scheme based on the "Use results data isolation targets" option. Then, if you are using a margin (e.g. 0.5) around each isolation window (as in the original SWATH manuscript and still modern SCIEX variable window methods), you would choose in the "Isolation width" field "Results with margin" and specify the margin.

Hope this helps. You can do it. We just don't help you create a method like this.

--Brendan

Hi Mike and Brendan,

Thanks for the ideas. It's help a lot.

The basic idea of mine is let the majority precursor range have smaller isolation windows based on RT. I hope it will improve both identification and quantification. Less complexity should archive it. I will have a detail method for this after I test Brendan's idea to make sure the machine and the software can do what i am going to do.

Thanks again!

Qi