Dear community,

I am working on modeling the structure of a protein for which no structural info is available. I have already generated a starting structural model with a tertiary structure prediction software. I also have DSS crosslinks from an XLMS experiment and I would like to use them to refine the starting model, i.e. as constraints.

So far, I have considered as constraints the unique crosslinks between two residues i and j (Lys), even if each crosslink can appear more than once in the measurement. The frequencies (for each crosslink, the frequency equals how many times the spectrum occurs) vary from 1 to 120 and they are not randomly distributed as a function of the crosslink separation j-i (see figure). My question is, can I interpret the high frequency crosslinks as those that form between residues that are (i) solvent accessible and (ii) thermodynamically stable? Viceversa, the low frequency crosslinks would be those involving residues that are in dynamical regions of the protein, in region that are not always solvent-accessible or both. By knowing this I could take a decision on whether I can give the constraints a weight that is proportional to the experimentally measured frequency of the crosslinks.

I know that there are many factors in a XLMS experiment and they might influence the frequency of the crosslinks. However, I do not have enough knowledge to understand whether these factors could be more important than thermodynamics and residue accessibility. I would appreciate to have your feedback.

Thanks!