Different dimethyl labels on different lysines within same peptide

[p-bell]

Hi,

Thanks again for developing these excellent tools.

I'm analysing dimethyl-labelled samples, and I'm running into issues with search space + some incorrect PSMs. Perhaps this is addressed by another feature but if not, I might be able to suggest a couple of solutions:

Problem: I'm labelling at the protein level looking for neo-N-termini, so have to include heavy or light mods for peptide n-terminus, in addition to dimethyl modified lysines (which are not cleaved by trypsin). I find that inclusion of heavy and light dimethyl variable mods on lysine / peptide n-terminus, increases the number of modified peptides to a huge amount, where number of possible dimethyl mods on K is set to >2. Add in semi- trypsin R digestion, plus oxidized methionine and deamidation on Q or N and it becomes a lengthy search on my workstation.

This is not the biggest problem though - by having these as variable mods, PSMs are detected that pass 1% FDR that contain lysines with a heavy label, and other lysines with a light label (within the same peptide). These are incorrect assignments, which likely interfere with the FDR calculation / thresholds applied. Is this a known issue?

Solutions?:

1. Is is possible (via an existing or new feature) to implement 'semi'-fixed modifications to help ID of dimethylated peptides? i.e. inheriting the fixed mods (eg alkylation on Cys), then having a +28K (light) or +34K (heavy) 'semi-fixed' mods... then true variable mods (oxidised Met etc) a level below these? This would likely reduce the search space considerably and reduce the risk of mis-identification of peptides with mixed labels.

2. Or alternatively perhaps a parameter for mass-shift searches that allows you to specify the mass shift based on the number of lysines (or any specific amino acid) in each peptide?

Thanks and best wishes, Pete