TMT based DDA Precursor Identification Benchmark

[Cajac102]

Hey everyone,

@RalfG came up with another way to test the precursor identifications of proteomics search engines. The main idea is to use TMT tags to indicate in each spectrum from which species it originated from:

And then this information can be used to calculate metrics such as FDR based on the species the search engine assigns to the spectra.
We already have fitting data from the LSMBO in Strasbourg with multiple TMT tagged organism mixtures, e.g. Arabidopsis-HeLa and Mouse-Arabidopsis-HeLa.

Current issues are removing the noise from the TMT channels, how to deal with shared peptides between species, and spectra with multiple TMT tags. Also, it might be too easy for search engines to distinguish between unrelated species (think human and flies), leading to almost similar performances across the search engines.

We are currently exploring the data more, if you have any feedback and aspects we have not yet thought about, feel free to share it here!

Cheers,
Caro