Modification not resolved

support
Modification not resolved Gereon  2018-08-22 03:30
 

Dear support team,

I tried to quantify a few petides with the DDA with MS1 filtering worklflow (Skyline 4.1.0.18169). Here I would like to quantify also two variable cysteine modifications (NEM and its heavy labeled variant NEMD5). As for some peptides there might be also combinations possible I tried to define both as structural (variable) modifications. The NEM modification is resolved pretty well, but the heavy one (NEMD5) is not recognized by skyline althought the corresponding peptide identification is present in library. Here, it is written that the modification (C[+130.07632] ) cannot be resolved. The theoretical shift is +130.079062.
The data (aquired by a QExactive plus) has been searched by Maxquant (version 1.6.1.0). I already tried to rename the modification both in skyline, the msms.txt and modifications.xml, but this did not solve the issue.
I'd be grateful for any help.

Thanks,

Gereon

 
 
Brendan MacLean responded:  2018-08-22 11:38

Hi Gereon,
That's an interesting one. A 2.7x10^-3 difference in m/z would be 2.7 ppm at 1000 m/z or 5.4 ppm at 500 m/z. Pretty close to what we might expect a modern HRMS instrument to detect.

Thanks for supplying all the files. I can see that the "NEM-D5" modification is defined in the modifications.xml as:

composition="H(2) C(6) N O(2) Hx(5)"

And in Skyline as:

H2H'5C6NO2

Which are exactly the same atomic compositions. So, I guess we need to dig into how they end up differing by so much. I will email Juergen to see if I can get the atomic masses he is using.

We will be way better off if the tools just agree on the atomic masses. The masses we use for this calculations are (taken from the code):

H = 1.007825035
H' = 2.014101779
C = 12.0
N = 14.003074
O = 15.99491463

Which match exactly with the ones in Unimod:

http://www.unimod.org/masses.html

You can do the calculation yourself in Excel, as I just did, and see that you will also end up with the theoretical mass from Skyline. So, I think we need to understand why MaxQuant gets a different mass. That is the crux of this issue.

Thanks for reporting it. I'll let you know what I find. If you have any way of dissecting the atomic masses used by MaxQuant, please add that information here.

--Brendan

 
Brendan MacLean responded:  2018-08-22 11:55

Note that a complete list of our atomic masses and notes on where they come from can be found in the code here:

https://github.com/ProteoWizard/pwiz/blob/master/pwiz_tools/Skyline/Util/BioMassCalc.cs#L380

 
Brendan MacLean responded:  2018-08-22 17:17

Nice quick response from Juergen that they use NIST exclusively as their source for:

H = 1.0078250321
H' = 2.014101778
C = 12.0
N = 14.0030740052
O = 15.9949146221

Code is here:
https://github.com/JurgenCox/compbio-base/blob/master/BaseLibS/Mol/ChemElements.cs

Original source is here:
https://physics.nist.gov/cgi-bin/Compositions/stand_alone.pl

We refer to this source as well, and these values should lead to the same modification mass out to 9 decimal places.

After a bit more digging, it appears that BiblioSpec (which has its own set of atomic masses for MaxQuant) until April 27 of this year had the wrong mass for deuterium (Hx). It had 2.01355321270, which does appear to be a value that is out there on the web (http://www.newworldencyclopedia.org/entry/Deuterium), but MaxQuant uses 2.014101778 (from NIST) and Skyline uses 2.014101779 (probably an older version of NIST).

So, if you switch to Skyline-daily, you should be able to rebuild your BiblioSpec library and continue working. Unfortunately, we have not gotten this fix into a Skyline 4.1 patch release and we are unlikely now to do so. It only affects deuterium labeling modifications in MaxQuant.

Here is the original support thread, which I am sure Kaipo would have found much more quickly, but he is away traveling.

https://skyline.ms/announcements/home/support/thread.view?rowId=36336

Thanks for the reminder of this issue.

--Brendan

 
Gereon responded:  2018-08-23 02:29

Dear Brendan,

thanks for taking care and doing a great job!
Using Skyline daily - as proposed - there is no more issue with resolving the heavily labeled modification.

Best regards,

Gereon