MSstatsPTM TMT / Label-free mixed input

[Klemens Fröhlich]

Dear MSstats Team,

We recently stumbled over the new MSstatsPTM and are very excited as we are increasingly doing more phospho analyses in our facility and would like to also use MSstats for phospho analyses.

Our standard approach usually is to do a phospho enrichment and measure the enriched phospho peptides directly (now with DIA). The flowthrough is collected and labeled with TMT and then also measured.

We therefore end up with a label-free phospho dataset, with protein level information stemming from TMT data.

Is it possible to analyse such data with MSstatsPTM, or are the only options:

label-free PTM input combined with label-free PROTEIN input 

or

TMT PTM input combined with TMT PROTEIN input

Again, we would have:

label-free PTM input combined with TMT PROTEIN input.

Best, Klemens

[Devon Kohler]

Hi Klemens,

On the method side this is definitely something that can be done with MSstatsPTM, however currently the implementation does not have the functionality to handle inputs from differing acquisition methods. I have added this as one of the next major feature enhancements to the package. If you have experiments you want to run through MSstatsPTM in the meantime, it is possible to do this using some of the underlying functions in the package along with the existing code in MSstats and MSstatsTMT. If you'd you like some guidance on how to do this let me know!

Best,

Devon

[Klemens Fröhlich]

Hi Devon,

We would be very interested in hearing how one could do this! We are also happy to share the results later here, so that others can have a look at how to do this.

Our initial concerns were about whether the statistical tests used for the phospho and the protein analysis make some underlying assumptions for both datasets (which in turn have to be the same). 

Can you please briefly comment on the way the analysis works, specifically whether the statistical analysis is done separately on phospho and protein data and only later the results are combined (in another model?) or whether you combine everything directly in a model.

Best, Klemens

[Klemens Fröhlich]

Hi Devon,

As I already mentioned we would be very interested in your input on how to do this!

Best Klemens