Hi Jean-Christophe,

Thank you for posting your question to the support board. You should take a look at the Panorama AutoQC pipeline if you haven't already done so. This pipeline lets you track the performance of your LC-MS system over time. Skyline documents containing results from system suitability samples are uploaded to a special type of folder on the Panorama server where you can see trends with standard deviations for various metrics, including retention time, mass error etc., in Levey Jennings plots that help you to quickly identify potential issues. Several labs have implemented the AutoQC pipeline for tracking system suitability. You can get more information about the pipeline on this page: https://panoramaweb.org/home/wiki-page.view?name=qc_with_panorama.

The Skyline and LabKey teams did a webinar recently on the "Best practices for evaluating LC-MS system suitability with Skyline and Panorama". This should be a good starting point. Here is the link: https://www.labkey.com/webinar/lc-ms-system-suitability-skyline-panorama/.

Please let us know if you have any questions.

Thanks,
-Vagisha

Dear Vagisha,

Thank you for yourfeedback. I tried to do exact what you show in your wikipage you sent me.
Unfortunately I was not able to setup the autoqc loader.

I send you a print screen of what I did, with a proxy error in the AutoQC configuration, and on this specific working station I use, we are able to go on the toll store. This means something in the config is not correct.

I am not sure, what I have to indicate in the "folder on Panorama".
Is it in our case:
INSELSPITAL-Clinical Chemistry
INSELSPITAL-Clinical Chemistry/test_acylcarnitine
Was unclear for me.

Or maybe there is another problem in the configuration of our autoQC loader?
Thank you in advance for your support.
Jean-Christophe

Jean-Christophe,
About the "folder on Panorama", in your case it is "INSELSPITAL - Clinical Chemistry/test_acylcarnitine"

The proxy error looks to be the same problem with your PCs proxy configuration that you reported in the other Support ticket (https://panoramaweb.org/home/support/announcements-thread.view?rowId=7828). The URLs which AutoQC is trying to reach are

https://panoramaweb.org/
https://panoramaweb.org/security/home/ensureLogin.view
https://panoramaweb.org/admin/home/healthCheck.view

These URLs will be accessing TCP/443 port on webserver.

Brian

Dear Brian,

Thank you for your help. I know it is exactly the same problem. I tried to explained it to our outsourced IT people.
In the hospital, there are so many firewalls, and proxy settings, it is really annoying. And this makes also very difficult to reach directly the right person to solve your problem.

It is completly crazy, since we have on two different pc's, similar settings, but one goes through and one is blocked.
I will transfer your infos to our IT, hoping they will be able to solve my problem.

Best, Jean-Christophe

Dear Vagisha,

Now everything is setup, our QC loader works, our intern proxy problems are solved. I try to follow the links you sent me on 2022-11-11 11:54:53 for my report I want to generate in Panorama.
I tried to find my metrics, to generate a table where I could see my CV for these different metrics, I red a lot of you wikipages, but unfortunately I was not able to get something similar as my excel sheet.
I took contact with labkey, since they organized a webinar, how o setup your lab with success. I thought they will be ablt to help me, but they refered me again to you, regarding my QC trend reports. (see attached file)

In my attachemnt there are also some open questions on our side, would it be possibel to answer them?

I am sorry for my incapacity to setup these QC trend reports, but I still need some help.

Kind regards
Jean-Christophe

Hi Jean-Christophe,

Glad to hear that you solved the proxy problems!

I can view your Word document that you originally attached, but not the PDF from your most recent update (in a variety of PDF viewers, it tells me that I need Acrobat Reader X, which has been end-of-lifed since 2015).

In your Word document, it looks like you have four samples and are analyzing twelve molecules, and tracking retention time, mass error, peak area, and peak height. There are four values for each metric, and two mean and CV values. Which sample(s) are you using to calculate those means and CVs?

We have just developed a "Trailing Mean" and "Trailing CV" plot type that will soon be available here on PanoramaWeb.org. It will visualize the CV and mean over the last N samples, where you can supply N. It does not export in a spreadsheet format yet, but perhaps it's similar to your request?

Thanks,
Josh

Good morning Josh,

Sorry that you were no able to read the PDF I sent you. I made two tif images and I send you also the excel sheet we are currently using to check our SST.
You can see which Skyline data are used for these different means and CVs.

As I already tried to explain in my mail to labkey, the goal would really be to be able to check the quality of the current measured SST (CV and STD of the different metrics we are interested in), to know if the sample analysis can be proceed or not.

And one additionnal question was, the possibility to re-integrate the peaks in panorama in case of wrong integration.

Thanks a lot for your so nice support.
Kind regards,
Jean-Christophe

Hi Jean-Christophe,

You can see the new Trailing Mean and Trailing CV plots in action on PanoramaWeb. Here's an example folder:

https://panoramaweb.org/home/Example Data/Quality Control/project-begin.view

Select them from the Plot Types drop down, and control how many samples get included in the mean and CV via the "Trailing Last" entry field, which defaults to 10.

As I noted, we don't yet have a mechanism to export the data to a spreadsheet, nor do we identify any of the calculated means or CVs as being outliers.

Looking at your spreadsheet, it looks like you are running four samples. You calculate the CV for the first three, and then for all four. I'm not seeing immediately how you define the color coding for the values, but it looks like perhaps 15% CV is your cutoff? And 5 ppm for mass accuracy?

Thanks,
Josh

Good morning Josh,

Regarding CV coloring, everything below 15% is green, orange above 15 till 15.5 and red above 15.5.
For the mass accuracy, everything within +- 5 ppm is green. between 5-10 ppm orange and red bellow 10 ppm.
We really do not need to export the data in spreadsheet, we should be able to check the SST before measuring study samples to be sure the system works corretly.

And in this particular case case I showed you, after 3 flush samples, 3 SSTs are measured, and these SSTS should show good mass accuracy and CV below 15%. The fourth SST is normally measured at the end of the sequence to have a also an in-sequence stability follow up of the system.

And regarding my question of re-integration in panorama, in case of wrong integration? is this possible?

Thank you a lot for the support, and I will check your example you sent me.
I wish you a nice start in the week and kind regards
JC

Thanks for explaining your thresholds. I'm surprised that you're defining such a narrow CV range for the orange color - 15 - 15.5. Is it common for you to see values in that range? We haven't yet implemented outliers definitions based on absolute values like this (and your mass accuracy cutoffs) but have been thinking about it for a while.

Sorry to miss your question about re-integration. We don't support edits like that directly in Panorama, as that would require duplicating a lot of Skyline's feature set. We are instead are focused on trying to make it easier and faster to iterate between Panorama and Skyline. In particular, consider trying the .skyp download format, which should be an easier approach than downloading the .sky.zip file and opening in Skyline.

https://www.labkey.org/Documentation/wiki-page.view?name=panoramaRevisionTrack

Thanks,
Josh

Good Monring Josh,

Our thresholds completly arbitrary. Regarding outliers that would be really a nice tool to have, or maybe have the possibility to add a dash line by 15 %, simply ticking a box.

I will try the .skyp file to check this question of re-integration. thank you for the tip.

Yesterday I could already test the trailing last, and it is nice that you have added such possibility. My question is, would you think to add the feature to be able to choose the samples we want to have a CV?

And one last question, I did a custom metric. I defined the Height, followed the wikipage you show for "define Custom Metrics". Unfortunately I am not able to show all Heights for each precursor, as for example for the mass accuracy. What can I changeto see all my precursor?
Or would it be possible to add the Height in default When configure QC metrics?

Sorry I have to add one question. I tried to import SST samples from a wiff file generated on a Qtrap 6500+. Unfortunately, we put all injections in one single wiff data set. When I set up the AutoQC folder and I put one wiff file he was not bale to recognize my SST samples. Should I change something in the config?

I really appreciate all your help
Have a nice day
Best, JC

How do you want to choose the samples to include in the CV calculation? Are you trying to filter based on the sample name? Or clicking to select a handful of samples in an ad-hoc way? Or something else?

Did you create a new SQL query to pull the height as a metric? What does it look like? Following the pattern you see in targetedms.QCMetric_peakArea should be a good starting point. I think you'd just need to make it pull the value of MaxHeight from PrecursorChromInfo instead of TotalArea.

I will refer your question about filtering for certain injections from a single WIFF dataset to Vagisha. I know this has been requested before but I'm not sure if it's been implemented.

Thanks,
Josh

Hi Jean-Christophe,

The "File Filter" option in the AutoQC Loader configuration only applies to file names, not the injection / replicate names within a wiff file. Sorry about that. The ability to filter which injections to import from a multi-injection wiff file was added to the Skyline command-line interface a while ago. But it is not yet available through AutoQC Loader. I have added it to the issues list here: https://skyline.ms/issues/home/issues/details.view?issueId=940. I will let you know when it is available in the next release.

Panorama will let you exclude replicates from QC plots and statistics calculation. When you click on a point in one of the QC plots you will see a pop-up box with the name of the replicate and an option to "Exclude replicate for all metrics". You can choose this option for all your non-SST replicates. I know this is not ideal once you are routinely running the AutoQC pipeline but I thought I would mention it while you are testing with existing files.

One thing I noticed in your screenshot is that you will have to change the selected "Instrument type" to SCIEX for AutoQC Loader to recognize wiff files.

Thanks,
Vagisha

Dear Josh,

That is a very good question, I do not know. But I can tell you what I unconsciously first tried to do. As you mentionned, to click on the wanted samples. but filtering with the name could also be an idea.

Back to the metrics, my custom metric, is at the end a peak area, that i do not want. I would need Height. As I told you I followed exactly what is explained in the Wikipage for adding custom metrics. I tried also to da what you explained abaove, choosing the "PrecursorChromInfo" but I got a red warning message (Add new metric_Height_2.tif). I did not really get how to customize really a new metric, to be honnest, there are too many possiblities for my possible understanding.

Adding Height as default in the QC metrics, would not be possible?
I will check also internally, maybe this height is not as important than the other metrics, and maybe we would be able to avoid it.

@Vagisha: Thank you so much for your so fast and late (1:40 AM in the US) repsonse. Since we know it, we will test it excluding what we do not want.

Best,
Jean-Christophe

Dear Panorama Team,

I have one new question regarding importing our data with waters instruments.
In attachment, a print screen of our sample List in MassLynx. You can see, our Raw File Name is the date, Molecule name and number increments.
Unfortunately in AutoQC i do not know how to fix the filter, to choose, our SST, our QC (ERNDIM, RECIPE), since they all have different names.
The filter is certainly only possible on the file name and not on meta data as Sample ID or file text?

Is it possible for example, to set 004-006 for SST, 015, 016 for ERNDIM and 019,020 for the Recipe?
Or do we have to do a complet adaptation of our sample lists?

One more thing, I noticed during the importatino of my wiff files, that in panorama some wiff files get and some get no imported.
And I think I understand why. We have a full scan method (our Flush method) at the end of our sample list in the same wiff file for cleaning our column.
It is also not imported in skyline. This leads to no import possibilites in panorama. Does I see it correctly?

Thank you in advance for oyur help and I wish you a nice day.
Kind regards,
JC

Hi JC,
Typically someone planning to use AutoQC would put some kind of identifier in the file names they are exporting that could be used for this filter, or direct them to be exported to a separate folder on disk from the main experiment files. If you have already acquired files named as in your screenshot:

..._001.raw
..._002.raw
..._003.raw
..._004.raw
..._005.raw
..._006.raw

I might still advise you to change the file names:

..._001.raw
..._002.raw
..._003.raw
..._004_SST.raw
..._005_SST.raw
..._006_SST.raw

And then use a simple filter like "Ends with" = "SST". Otherwise, you can create a regular expression to select just the files you know you want to include, but this would be only a one-time filter, e.g. "_(004|005|006|015|016|019|020).raw" should do what you have requested, but it is a very specific and brittle implementation that may serve you in this one case.

I highly recommend this Regular Expression tester for helping to build regular expressions that match what you are looking for.

https://regex101.com/

--Brendan

Dear all,

6 month ago I asked if single injections could be implemented in AutoQC from Wiff files, and Vagisha told it was not possible yet.

"The "File Filter" option in the AutoQC Loader configuration only applies to file names, not the injection / replicate names within a wiff file. Sorry about that. The ability to filter which injections to import from a multi-injection wiff file was added to the Skyline command-line interface a while ago. But it is not yet available through AutoQC Loader. I have added it to the issues list here: https://skyline.ms/issues/home/issues/details.view?issueId=940. I will let you know when it is available in the next release."

I was wondering if you could give me a short update about that. Could it be implemented?

In advance thank you for your help and support. I wish you all a nice summer and happy to hear from you.
Kind regards
Jean-Christophe

Dear Jean-Christophe,

Thank you for following up. We haven't yet implemented the option for sample file filtering from multi-injection Wiff files in AutoQC Loader. I just added your email address to the notify list for https://skyline.ms/issues/home/issues/details.view?issueId=940. You will receive a message when we make progress on this.

Best regards,
Vagisha