| |
| Nick Shulman responded: |
2018-06-19 10:44 |
That might be a bug in Skyline. It looks like Skyline is getting stuck right at the point where Skyline has finished importing your iRT peptides, and is trying to figure out what slope and intercept to use for mapping from iRT to retention time.
Is this a problem with all of your .raw files, or just that first one? Have you tried this on other computers and does it work?
Can you send us your files?
In Skyline, you can use the menu item:
File > Share > (Complete)
to create a .zip file containing your Skyline document and supporting files, including iRT database.
You can upload that .zip file here:
https://skyline.ms/files.url
You should also send us at least one of the raw files that you are trying to extract chromatograms from.
|
| |
| Brendan MacLean responded: |
2018-06-19 11:16 |
Hi Josh,
Can you right click on CPU graphs and choose Change graphs to > Logical processors?
Not that it explains such a long pause at this point, but there will be some kind of pause after 40% while, as Nick explains, Skyline recalculates all of the extraction parameters for the nearly 1 million transitions in your document. The first 40% is extracting full-gradient chromatograms for your iRT standards. Though, if it is showing 41%, I suspect it is already past this calculation, which the progress UI also indicates with the gray vertical bar at 5 minutes. Though, I have myself recently been importing files with over 2 million transitions on a desktop i7 system with 16 GB of RAM. So, there is definitely something strange here.
Can you provide a screenshot of your Transition Settings - Full-Scan tab?
Sending us the files as Nick suggests would also help tremendously.
--Brendan
|
| |
| Josh Baeza responded: |
2018-06-19 13:38 |
Hi Nick and Brendan,
I uploaded my skyline document along with a raw file to test. Thanks for the support.
I have tried to import data into skyline on my laptop (Dell XPS 15, same Skyline daily version) and it was also very slow. I don't think it even reached the 40% mark after 8 hours.
I went ahead and attached the screen shots of the CPU processors and there doesn't seem to be anything abnormal.
Thanks again,
Josh
|
|
| |
| Nick Shulman responded: |
2018-06-19 14:07 |
Thanks for sending us your files.
The problem is that you have Overlap Deconvolution turned on.
Overlap deconvolution is an algorithm for dealing with overlapping DIA isolation windows. We have been meaning to remove this feature from Skyline, because the recommended way of dealing with such data is to use MSConvert to create .mzML file that has been deconvoluted.
The way that we implemented in Skyline really cannot handle a large number of peptides.
Anyway, you probably never intended to turn this on.
You should go to:
Settings > Transition Settings > Full Scan > MS / MS filtering Isolation Scheme > Edit Current
and then change "Deconvolution" to "None" |
| |
| Josh Baeza responded: |
2018-06-20 05:27 |
Thanks Nick.
I acquired my data using overlapping windows and I thought the raw files would be demultiplexed in Skyline. This was my approach because I couldn't get MSConvert to demultiplex my files. I actually met with you at ASMS and I mentioned the problems I was having with MSConvert. Well I tried to run MSConvert again today on a different system and it also fails. I've attached the screen shot and the failed mzML file.
Sorry to change the topic of this conversation. Let me know if you would prefer I begin a different thread instead.
Thanks,
Josh |
|
| |
| Nick Shulman responded: |
2018-06-20 14:29 |
I will try to figure out what is going wrong with MSconvert.
The error I am seeing is:
Failed - System.Exception: SpectrumToIndices() Number of demultiplexing windows changed. Window boundary tolerance may be set too low.
at pwiz.CLI.msdata.MSDataFile.write(MSData msd, String filename, WriteConfig config, IterationListenerRegistry iterationListenerRegistry)
at MSConvertGUI.MainLogic.processFile(String filename, Config config, ReaderList readers)
at MSConvertGUI.MainLogic.Go(Config config)
I will ask around and see if I can figure out who can fix this. |
| |
| atkeller responded: |
2018-06-22 18:23 |
Hi Josh,
It looks like within each DIA cycle, the window centers are offset by 23 m/z instead of 24 m/z. This is causing msconvert to fail during demultplexing because the windows won't align to a nice repeating grid. Unfortunately, we can't demultiplex the data but you can still search it without any demultiplexing by setting "Deconvolution" to "None", as Brendan described above. You won't get the improved selectivity benefits of demultiplexing, but hopefully this will still let you salvage the data.
There's actually a feature in Skyline that will generate an optimized set of DIA windows as a table of Start m/z and End m/z values with the option to apply overlapping. This is meant to help guarantee that a chosen windowing scheme can be demultiplexed.
In a blank Skyline document, go to
Settings > Transition Settings > Full Scan > MS / MS filtering
and then set "Acquisition method" to "DIA".
Next, within "MS / MS filtering" go to Isolation scheme > Add... > Prespecified isolation windows > Calculate...
This will open a "Calculate Isolation Scheme" dialog where you just need to specify "Start m/z", "End m/z", set "Deconvolution" to "Overlap", and check "Optimize window placement", then select "OK" in order to generate a valid overlapping scheme. The resulting table can be copy-pasted from Skyline in order to transfer it to your method editor.
Let me know if this is helpful, or if you were trying to do something different. Did you know about the feature for generating overlapping windows within Skyline? Maybe this is something we could make easier for users to find.
Thanks,
Austin |
| |
|
|
skyline_slow.pptx
20180525_Baeza_circadian_dia_zt-01-37.mzML