Feature Request: Accounting for MS3 data for cleavable Crosslinkers

[atroel01]

Hi xiSearch team, this is a fantastic suits that you have and very much exceed expectations, even for a paid commerical product.

Requested feature: MS3 integration.

Inclusion of peptide spectral matches from MS3 fragmentation into the FDR algorithm would be valuable, especailly for cleavable crosslinkers. Not being familiar with the precise code of how FDR is calculated in xisearch, but from basic principles, MS3 data of linked peptides clearly adds an additional validation to the MS2 annotated spectra used currently for matching. This approach is implemented in XlinkX/Metamorpheus/MSAnnika, though each of them have limitations that xisearch addresses, especially around the control over FDR and crosslinking sites.

Could the addition of MS3 data be implemented as an option in xiSearch/ xiFDR, or at least included down stream, so it can be visible during use of the XiView, showing alignments of the MS3 fragmentation inverted below the chimeric MS2 spectra or overlayed over the chimeric MS2 spectrum?

[lutzfischer]

While the I can see that if you actually acquired MS3 data it would be nice to make use of them, however we don't use any MS3 data in our lab. Implementing that would require a rather large change/addition to the code. So I don't think it will happen any time soon.
And to be honest I think acquiring high resolution (e.g. >=60K) MS2 spectra with stepped collision energy is probably the better strategy when using xiSEARCH.

[atroel01]

That's good information thanks
I had attempted to generate 'ms2.5' spectra by combing ms2/ms3 hybrid spectra (using python and openMS) to feed into xisearch, as a way to boost ions useful for sequencing. It was an interesting exercise but not sure it really helps much in my data sets.

I appreciate the tip on acquisition settings, is 60k significantly better than 30k? Is there a reference for that?

[lutzfischer]

with 60k with a well calibrated instrument (or recalibration) we search sometimes with as low as <=5ppm MS2 here and that has quite a large impact on the search result.

[atroel01]

Thanks Lutz,
I came across the paper https://pubs.acs.org/doi/10.1021/acs.analchem.1c05266, where you and the team demonstrate the interesting value of MS2 only approaches, and indeed in my testing the stepped HCD MS2 approach seems to yield better and clearer matches to MS -cleavable spectra.
Once again, great piece of software and i look forward to future iterations.