Detection Z score and Q Value.

support
Detection Z score and Q Value. anirudhkashyap511  2024-07-30 12:53
 

Hi Skyline Team,
I want to run a Quality check on peaks that have been identified by my mprophet model trained on second best peaks. Could you guide me on what Z score and Q value I need to keep? Thanks!

 
 
Nick Shulman responded:  2024-07-30 13:03
I am not sure whether you can trust anything about a peak scoring model that has been trained on second best peaks.
I gave a bit of an explanation for that here:
https://skyline.ms/announcements/home/support/thread.view?rowId=51479

If you send us your data we might be able to come up with better advice.
In Skyline you can use the menu item:
File > Share
to create a .zip file containing your Skyline document and supporting files.
The Share Document dialog gives you the option to include the mass spec data files or you could send them to us separately.
Files which are less than 50MB can be attached to these support request.
For larger files, you can always zip everything up and upload them here:
https://skyline.ms/files.url
-- Nick
 
anirudhkashyap511 responded:  2024-07-30 13:36
Hi, Please let me know if you can open this. Although the peaks don't look good, for the ones that look good. I believe Skyline did a good job at picking the right peaks. Wanted to know if there was a way for me to do quality check. I'm already looking at the Intensity, Area, Background. Thought I will incorporate the Z score and the Q values as well.
 
Nick Shulman responded:  2024-07-30 14:32
No, that's not the correct file.
The file that you should send us will have the filename extension ".sky.zip".
That's a .skyd file which we cannot read without the rest of the files that will be in the .sky.zip that you get when you do "File > Share".
-- Nick
 
anirudhkashyap511 responded:  2024-07-30 14:38
I'm sorry for not reading your instructions correctly before. Here it is! Thanks again.
 
Nick Shulman responded:  2024-07-30 16:08
Thank you for sending that Skyline document.

I see that your document has 57 peptides, 65 precursors and 258 transitions.
How did you choose the transitions for each peptide?
It looks like a lot of your transition chromatograms have some interference in them. That is, the peak shape is not a smooth Gaussian and is a different shape than the chromatograms for the other transitions for the same precursor.
Sometimes three transitions is enough to choose the correct peak, but, if some of the transitions might have interference, it would be a lot better to be able to look at 6 transitions instead.

How did you choose which transitions to monitor for each peptide?
It looks like your document used to have a spectral library, but does not have one anymore.
It's often helpful to have a spectral library when choosing peaks, because the spectral library tells Skyline the expected ratio of transition peak areas.
If you do have a spectral library, I would recommend going to "Settings > Peptide Settings > Library" and add the library to the document.
After you have added the library, you can tell Skyline to pick peaks again by going to "Refine > Reintegrate" and choose "Default" in the Peak Scoring Model dropdown and check the checkbox "Overwrite manual integration".

If you do not have a spectral library you can use Koina to create one.
Koina uses artificial intelligence to predict which fragment ions will be the most intense.
I created a library using Koina and I have attached a Skyline document that uses it.
One thing that you can do once you have the Koina library is to look at the "Peak Area > Replicate Comparison" window and compare your observed peak areas to the predicted peak areas.
The best-looking peptide in your document is AALSGANVLTLIEK which I have attached a screenshot of. The observed peak areas agree fairly well with Koina's prediction. I did not see any other good-looking peptides in the document.

-- Nick