LFQ in Human Plasma: Benchmarking with a high dynamic range multispecies dataset

[rodvrees]

aim of the new module

This module proposal is based on the following preprint:
Ute Distler, Han Yoo, Oliver Kardell et al. Multicenter Evaluation of Label-Free Quantification in Human Plasma: Benchmarking with a High Dynamic Range Multispecies Sample Set, 18 December 2024, PREPRINT (Version 1) available at Research Square [https://doi.org/10.21203/rs.3.rs-5618718/v1]

1. what is its aim?

Benchmarking LFQ accuracy for high dynamic ranges, similar to those found in biomarker discovery studies using plasma proteomics.

2. what is its added value compared to already available modules?

The data is a three species benchmark dataset, similarly to currently available modules. However, there are some noteworthy differences:
Instead of HeLa cell digests, the human component in the three species sample is human plasma, completely changing the sample matrix. The spike-in of tryptic digests of yeast and E. coli proteomes (which are at much lower concentrations compared to the spike-ins from the data we use in the current modules) simulates the high protein dynamic range of human plasma and the low abundance of potential biomarker candidates.

full description of the new module

1. What file would be available for download on the module website?
JPST003358 (https://repository.jpostdb.org/preview/7939445236704f290b03ad, Access key: 4768)

There's a lot of files here, because the data was acquired on various instruments. Which subset we choose can be discussed. I suggest the timsTOF, since to my knowledge it's a popular instrument for plasma proteomics.
There is also both DIA and DDA data, again this can be discussed, but DIA has become the prevailing method for biomarker discovery in my experience.
Lastly the dataset also contains different levels of spike-ins, similar to the current HYE datasets used in the current modules. Again, the specific spike-in quantity can be discussed, or we can come up with an approach to use all different spike-in levels. 

2. What would users need to do?
Similarly as already existing modules. However, since the aim of the dataset is to create a benchmark set for LFQ benchmarking for biomarkers in human plasma, we might consider evaluating on the protein level, instead of the PSM level. To be discussed.

3. What would be the type/format of files used as input for the module?
Depends on the above. 
4. What metrics would be calculated and how?
Depends on the above. In the simplest case, the same metrics as in already existing modules. 
For the in-depth visualizations, there are some really interesting plots in the manuscript from which we can take inspiration. I particularly like the idea of the dynamic range visualization (Fig 3.), but probably cannot be used for the overview of all runs. 

5. Propose a plot for metrics visualisation
Depends on the above.

potential reviewers

TBD. The authors of the manuscript would be the most obvious choice.

Will you be able to work on the implementation (coding) yourself, with additional help from the ProteoBench maintainers?

• yes
• no

any other information

No response