Dear Skyline team,
thank you for providing us with the possibility to get numbers representing protein abundance per replicate directly from Skyline, which is also supposed to be part of the next main Skyline release. I intend to use this option rather frequently in my research and would like to understand it a bit better. I am mainly interested in the normalization case against heavy (each light signal must have a heavy counterpart to be considered).
In the release notes, you state that this protein abundance value is the same that the Group Comparison (GC) framework uses. I thought that, as default, on protein level GC sums up all transition areas that belong to the same protein (also across different peptides) that are quantitative and have a heavy counterpart and then normalizes this sum to the sum of heavy signals. This is slightly different compared with calculating normalized ratios transition by transition and then averaging those ratios, since strong signals dominate the sums of values, as previously described on several occasions. However, the explanation in the reporting window for the "Protein Abundance" feature now states that "The Protein Abundance is calculated by taking the average of the normalized areas of all of the Transitions under the Protein." This now sounds to me exactly as the latter option to do it and the question is, whether protein abundance numbers really are exactly what the GC framework uses?
Thanks a lot,
Roman
|
| |
| Nick Shulman responded: |
2020-10-12 09:12 |
Yes, the protein abundance number is exactly what the group comparison uses.
If you are doing ratio to heavy, then that number is calculated by taking the sum of the light transition areas across all of the peptides and dividing by the sum of the heavy transition areas.
However, transitions are only included in these sums if the matching transition is present and has a valid peak area on the other side of the dividing line.
-- Nick |
| |
| roman sakson responded: |
2020-10-12 09:38 |
Hi Nick,
ok, thank you! I guess I got a bit confused by the "taking the average of the normalized areas of all of the Transitions" statement in the report help but I am happy to read that it is still the sum.
Unfortunately, I also have a practical question: I created a document grid report to extract the protein abundance values in a document that contains only one protein, consisting of one peptide with a heavy and a light precursor present. I have 6 runs grouped into two conditions. For GC, I can normalize to heavy or not and everything works as usual. However, if I go to Peptide Settings and set "Ratio to Heavy" as the normalization for quantification, all my protein abundance values become #N/A. For no normalization or equalize to median, everything works and I get reasonable values. Do you have an idea what might be wrong and how I can get abundances normalized to heavy? I have also observed this in other Skyline documents (I am using the latest daily release). The document is shared below.
Roman |
|
| |
| Nick Shulman responded: |
2020-10-12 10:12 |
Oops. I wrote the "if" statement backwards. The behavior was supposed to be the opposite: The protein abundance should be #N/A if there are missing transitions, _unless_ the normalization method is Ratio to Heavy.
I will fix this. This fix will probably appear in the first Skyline-Daily which comes out after we release Skyline 20.2, which will probably be later this week.
Thank you for reporting this bug.
-- Nick |
| |
| roman sakson responded: |
2020-10-12 10:20 |
Sure, I am happy to help!
Thank you,
Roman |
| |
| roman sakson responded: |
2021-08-15 16:35 |
Hi Nick,
I decided to reopen this older thread, since I have observed some behavior of the protein abundance feature that I felt unsure about, while using the latest daily-release.
I have a heavy and a light precursor area for a protein (only one peptide in the document, two replicates). I have normal signals in one replicate but slightly truncated ones in the other ("Fibroblasts_2" truncated accordingly to Skyline). I am using "None" as Normalization method under Peptide settings, and I am getting #N/A for the protein abundance in the respective truncated replicate. Confusingly, as soon as I remove fragment ions y8+ and y5+ (both truncated in light) but keep y8++ (not truncated in light but truncated in heavy), I am suddenly getting a protein abundance value. I am getting the same when I keep all the transitions but remove the other, non-truncated replicate instead ("Fibroblasts_1"). Might that be some kind of a bug?
I am attaching the document.
Thank you in advance,
Roman |
|
| |
| Nick Shulman responded: |
2021-08-15 17:18 |
When your normalization method is "None", then Skyline only uses the light peak areas to calculate Protein Abundance. Skyline excludes the heavy stuff because the Internal Standard at "Settings > Peptide Settings > Modifications" is "heavy".
If you changed that Internal Standard to "None", then Skyline would calculate protein abundance by summing both the light and heavy transitions.
Yes, this seems to all be working the way it's supposed to.
When the internal standard is "heavy" and the normalization method is "none", the heavy transitions end up being completely ignored.
-- Nick |
| |
| roman sakson responded: |
2021-08-15 17:46 |
Hi Nick,
ok, thank you for the fast response, this explains the value I get after deleting light truncated transitions then! Are truncated areas generally being ignored by Skyline, even if marked quantitative, and how is the group comparison module treating those if normalization is set to ratio to heavy?
However, I am afraid that I still do not get why I am seeing #N/A for the "Fibroblasts_2" replicate as long as Fibroblasts_1 is present in the document but I am immediately getting the correct abundance for Fibroblasts_2 as soon as I delete the Fibroblasts_1 replicate from the document. I do not have to delete any transitions then and I am also not changing any peptide settings, just remove another replicate. Should not protein abundance values across different replicates be independent?
Thank you,
Roman |
| |
| Nick Shulman responded: |
2021-08-15 18:41 |
Skyline looks at all of the replicates in your document in order to decide which transitions are "required" for the Protein Abundance calculation.
If every one of the replicates is missing the value for a particular transition (either because the peak is missing, because you did "remove peak", or because the peak is truncated), then that transition is not required, and Skyline ignores it.
But, if at least one replicate has a value for that transition, then that transition is required, and every replicate that is missing that value will have a Protein Abundance which is #N/A.
The reason that Skyline does this is so that it will make sense to compare protein abundances across replicates. If the abundance was calculated using different sets of transitions in different replicates, then it wouldn't make sense to compare those numbers across replicates.
Once you get down to just one replicate in your document, any missing values will be seen as not required, and you will get a Protein Abundance value so long as at least one transition has a valid peak area.
-- Nick |
| |
| roman sakson responded: |
2021-08-17 17:29 |
Sorry for my late response and thank you for making all of this very clear now!
I am afraid that I have one more question regarding ratio to heavy quantification in my aforementioned document (simple precursor ratios NOT ticked). I have only one peptide there and for the Fibroblasts_1 without truncated signals I am getting exactly the ratio what I would expect (ratio of sums of light and heavy transition areas). For Fibroblasts_2 it is a bit more tricky, since I have only one light and heavy transition without truncation but those happened to be different ions, so I am not quite sure how Skyline calculates the ratio there but it is very close to what Skyline reports to be the "total ratio" for the only peptide in the targets table. Finally, when I tick the simple precursor ratios box under peptide settings, both my protein abundance values change quite substantially (going from above one to below one) and I am not sure what is going on.
Thank you in advance,
Roman |
| |
|
|
201012_20-65_Skyline_Support.sky.zip