MSstat output explanation of file "TestingResults"

82 views
Skip to first unread message

Melissa Vorster

unread,
Sep 15, 2022, 6:32:35 AM9/15/22
to MSstats
Good day

I am using the external tool of MSstat on Skyline. Could someone perhaps help with what the values in these columns mean and how they have been calculated in the TestingResults.csv that is returned when doing group comparisons?

Kind regards
Melissa Vorster

Mateusz Staniak

unread,
Sep 15, 2022, 6:59:49 AM9/15/22
to MSstats
Hi,


could you please attach a piece of this file or paste column names here?

Kind regards,
Mateusz

Melissa Vorster

unread,
Sep 15, 2022, 8:51:07 AM9/15/22
to MSstats
Here are the columns that I am referring to:

TestingResults.csv columns.PNG

Kind regards
Melissa

Mateusz Staniak

unread,
Sep 15, 2022, 9:13:56 AM9/15/22
to MSstats
Hi,

Protein: protein ID
Label: names of Conditions selected for comparisons separated by "vs"
log2FC: log-fold change between conditions listed in the "Label" column estimated by MSstats [basically ANOVA]
SE: standard error for estimated log2FC
Tvalue: test statistic for significance of log2FC 
DF: degrees of freedom of the distribution of test statistics
pvalue: p-value calculated based on the T-statistic
adj.pvalue: p-value adjusted for multiple comparisons using Benjamini-Hochberg method
issue: lists possible issues such as MSstats not being able to produce a p-value due to lack of biological replicates
MissingPercentage and ImputationPercentage: percentage of missing values and imputed values, respectively, in feature-level data used to produce the protein-level data that was then used to produce these testing results

Hope this helps, don't hesitate to let us know if you need additional information.

Kind regards,
Mateusz Staniak

Melissa Vorster

unread,
Sep 16, 2022, 1:53:02 AM9/16/22
to MSstats
Great, thank you very much.

Could you share the formula by which the log2FC has been calculated?
And how is the DF calculated?

--Melissa

Mateusz Staniak

unread,
Sep 22, 2022, 7:20:19 PM9/22/22
to MSstats
Hi,

it's a difference between estimated Condition effects.
DF depend on the model (it's either regular ANOVA or a model with random effects). I'm working on updating our documentation to show the details.

Kind regards,
Mateusz

Melissa Vorster

unread,
Sep 23, 2022, 4:32:59 AM9/23/22
to MSstats
Hi thank you for replying. For the log2foldchange, I am wondering which quantities it uses between the different condition effects? Is it protein abundance or intensity values that are used?
(Thank you will look at the link)

Would be great to be notified when the documentation is ready :)

Kind regards
Melissa

Mateusz Staniak

unread,
Sep 23, 2022, 4:36:57 AM9/23/22
to MSstats
MSstats summarizes feature-level intensities to abundance estimates (dataProcess). So when we test, there is a single abundance values per protein and run.


Kind regards
Mateusz

Melissa Vorster

unread,
Sep 23, 2022, 4:38:54 AM9/23/22
to MSstats
Ah, I see -  thank you. Could the code for this part be shared? What is the input (format) for this section?

Mateusz Staniak

unread,
Sep 23, 2022, 4:43:13 AM9/23/22
to MSstats
Hi,

for input format and processing steps, this manual is a good reference: https://msstats.org/wp-content/uploads/2020/02/MSstats_v3.18.1_manual_2020Feb26-v2.pdf
Some of it is outdated, but only in terms of some parameter options that we removed or names that we changed. The workflow, inputs etc are still the same.

Kind regards,
Mateusz Staniak

Mateusz Staniak

unread,
Sep 23, 2022, 6:01:54 AM9/23/22
to MSstats
Hi,


on a second thought, there is no need to wait for documentation update (which may take a while), the models are described here: https://docs.lib.purdue.edu/cgi/viewcontent.cgi?article=1743&context=open_access_dissertations

Kind regards,
Mateusz Staniak

Melissa Vorster

unread,
Sep 23, 2022, 9:35:40 AM9/23/22
to MSstats
Great, will give it a look. Thank you.

Kind Regards
Melissa

Reply all
Reply to author
Forward
0 new messages