Reporter isotopic distribution for TMT

[carlomartins]

Hi Phil,
I am starting to use PAW pipeline in my TMT studies. The transparency of the workflow and the flexibility is awesome! Although I am not acquainted with Python scripts, I think it is possible to understand what is happening in each step.
However I could not find where to add the correction factors for TMT reporter ions.

Thanks,
Carlo

[pwilmart]

Hi Carlo,
I do not have a mechanism for correcting for label purity. This limits the dynamic range some and that is not ideal. We mostly do SPS MS3 data in our core so that helps some for the dynamic range. Most correction methods are attempted at the PSM level. There can be missing data in the individual scans and I am not sure the usual linear algebra methods can tolerate that very well. It probably makes sense to do corrections at higher aggregation levels (like proteins) where there is much less missing data and noise. It needs to be done before any normalizations, so it would take some thought. I also think that fitting a sum of theoretical isotopic patterns to the measured data would be safer that doing the matrix inversion to get abundances from the measurements given the correction factors. It might sound like 6 of one versus half a dozen, but I think there is a fundamental difference.

All of this would take some serious effort, so I have no idea when I might get around to it. I have been thinking about it. I just live with fold changes of up to maybe 10-fold being the maximum that isobaric labeling can do. I think most situations are not negatively impacted. In the rare situations where one really is trying to NOT detect any signal for some proteins, those may not be the best experiments for TMT.

I have been thinking about writing a blog entry to discuss the pipeline design choices in more detail. There are also some common use cases that require some file management to make work that I should describe. A single step (black box) is easier than 5 steps until the use case for the one-step does not work for what you need to do. Multiple steps offers flexibility, but maybe at the price of some confusion.

Thanks for the suggestion on the correction factors. Please let me know if you have any other questions.
Cheers,
Phil

[carlomartins]

Thank you for your kind reply.
I will be eagerly waiting for your blog entry about file management, as I am very interested in data analysis for quantitative proteomics. Your pipeline is a valuable addition to the short arsenal of user friendly softwares to that goal.

Cheers,
Carlo