Dear specL maintainers,
I think I discovered a bug in specL. In short, I prepared a Mascot search including two N-terminal peptide modifications:
- dimethylation (N-term)
- Acetylation
after that I created a spectral library from these search results using the bibliospec implementation in skyline-daily and based on this I created a assay library for Spectronaut using specL. When importing the lib I all of a sudden get asked to "explain" a huge amount of modifications to Spectronaut (see screenshot). To me it looks like the following happend. You use the following to write modified peptide sequences:
X[+|-Int]XX[+|-Int]XXX with X being any amino acid in single letter code. While this might work for internal modifications, it creates confusion for the terminal modifications. A simple example: Mascot ident. the peptide to be N-term. dimethylated, specL expresses this as:
X[+28]XXXXXXX
But Spectronaut now needs to know if the peptide was modified on the side chain or on the alpha amine (N-term of peptide) and X[+28]XXXX is not clear in this respect.
Could somebody please look into this?
I had a look into the .blib files and saw something really sad: The BiblioSpec people already do not handle the N-term. modifications correctly. Here is the head of the Modifications table in the SQLlite dump (.blib file):
> head(my_Modifications)
Source: query [?? x 4]
Database: sqlite 3.11.1 [joined_Gluc_control.blib]
id RefSpectraID position mass
<int> <int> <int> <dbl>
1 1 2 3 57.02146
2 2 2 6 28.03130
3 3 6 3 28.03130
4 4 6 10 28.03130
5 5 9 13 28.03130
6 6 9 17 28.03130
The modifications is just an integer (Da units). No link to UNIMOD IDs. Really poor, since the Mascot server used the org. UNIMOD definitions. di-methylation is UNIMOD acc 36 and it comes with different specificities:
http://www.unimod.org/modifications_view.php?editid1=36
The Bibliospec people do not catch this information in any way but flatten it to an integer :-(