Downstream analysis of MaxQuant phospho PSMs

[dh2305]

Dear all,

thank you for providing a TIMS-specific rescoring platform.
I tried this tool on a MaxQuant Phospho dataset, and I was wondering what one is best supposed to do with the output.

I can assess I have increased PSM discovery but as there are no intensity attributes in the peptide output of mokapot I cannot directly continue from here.

How can I actually use the main output (psms.tsv) to make use of the increased PSM identification in the set FDR?

I was so far unable to re-import this into MaxQuant (or Fragpipe) for partial processing. I would appreciate hints at how to end up at tables akin to the original maxquant output regarding modified peptides or Phospho(ST) sites.

Best,
Dominik

[RalfG]

Hi Dominik,

Thanks for reaching out!

How can I actually use the main output (psms.tsv) to make use of the increased PSM identification in the set FDR?

That would depend on what you would like to achieve downstream. Are you interested in identified phosphosites, or in quantification info? For the former, you could filter the output on all PSMs that have a phosphorylation and pass the FDR threshold. The latter would require a bit more work, as the identifications would need to be either requantified with a tool such as FlashLFQ, or the PSM scores from MaxQuant need to be updated in the entire MaxQuant output.

We'd be happy to help you get started in any case.

Best,
Ralf

[dh2305]

Hi Ralf,

thanks for coming back to me this quickly!

While we can definitely see utilization in questions of identification, in this case, we are interested in quantification/output formats akin to conventional LFQ workflows.

We are using MaxQuant in auxiliary/complementing our Frapipe-based ddaPASEF downstream workflows, especially in phosphoproteomics, to gain an additional layer of confidence by not exclusively relying on one (site allocation) algorithm.

As MQ does not make use of rescoring (in contrast to the Perculator incorporation in FragPipe or, e.g. whatever DIA-NN is doing), we thought to explore the options of improving coverage with tims2rescore for our ddaPASEF applications on this dataset.

So, to make use of the expanded PSM matrix and be able to benchmark performance, we would ideally need a Phospho (ST) quantification table akin to the original MQ output and the complementary Fragpipe dataset. As I can only suppose that site-focused processing will be an additional layer of figuring this out, we would be happy to assess performance on (phospho)peptide level only as well. Here I don't know if, e.g. MSstats (PTMstats) will help us derive quantification tables from the tims2rescore PSM table output. My experience has so far not exceeded partial processing inside the suites for these cases.

Best,
dominik