DIA Profile vs Centroided data

DIA Profile vs Centroided data boycea  2021-10-28

Hello Skyline Support Team,

I am currently acquiring my proteomics DIA data on the Thermo Orbitrap Exploris 480 using profile spectra because we were told from a colleague that this gives us more accurate data. Since we set it up this method, we have seen mixed opinions on whether centroided or profile data is preferred for DIA, and since the new Skyline daily update allows switching between both, we were wondering which of those the Skyline Team recommends. Another instrument we use for DIA acquisition is the Thermo Fusion Lumos. Thanks!


Nick Shulman responded:  2021-10-29
We recommend that you choose "Centroided" as the Precursor and Product Mass Analyzers at "Settings > Transition Settings > Full Scan".
When you choose that setting, if the .raw file contains profile data, Skyline will ask Thermo software to centroid the profile data.
Thermo's algorithm for centroiding profile data does a better job of figuring out what the m/z and area of a peak in a spectrum is.
If you instead choose a mass analyzer such as "Orbitrap", Skyline will extract chromatograms from profile data by summing the intensities across a m/z range, and the width of the m/z channel that Skyline is summing across will be approximately equal to the resolution of the peak.

We do not have an opinion as to whether you should tell the mass spectrometer to collect profile or centroided data. I know your raw files will be smaller if you tell the mass spectrometer to only keep the centroided data. I have heard that sometimes human beings are able to look at profile data and see things that might be going wrong in the mass spectrometer which would not be obvious if they only had the centroided data to look at.
-- Nick
boycea responded:  2021-10-29
Hi Nick,

Thank you for the reply. You mentioned skyline works with the .raw file that contains the profile data. We have been using Proteowizard to convert the .raw files to .mzML files, is Skyline able to directly handle the multiplexed files with overlapping windows now? What about when we merge multiple multiplexed files with overlapping windows into one like we do for our GPF injections, can skyline view that as .raw, or preferred as .mzML?


Nick Shulman responded:  2021-10-29
You should do your demultiplexing in msconvert if you have overlapping isolation windows. Skyline does not know how to do the demultiplexing that msconvert does.

When you do the demultiplexing in msconvert, my understanding is that you should also have the peakPicking filter applied. The "peakPicking" filter in msconvert performs centroiding on the spectra. I believe the demultiplexing algorithm works a lot better on centroided data compared to profile.
-- Nick
Brendan MacLean responded:  2021-10-31
Agreed. You want to perform demultiplexing with msconvert peakPicking or MSConvert GUI with "Peak Picking" filter and "Vendor" algorithm.

When you do not have multiplexed MS/MS, then you would import the raw data into Skyline choosing the "Centroided" mass analyzer. This has become common enough in our lab that researchers often acquire centroided only data, because the raw data files are smaller, and the centroiding is identical to what you get from a profile mode raw data file. You are keeping more information with the profile mode file, and that can be informative when you really want to dig into the details of why centroids end up where they are, but for large-scale DIA, it can be a lot of extra disk space for this benefit.

For demultiplexed data, it is not practical to have both spectrum types passed through the demultiplexing algorithm.