bigDIANNtoMSStatsFormat annotation file for multiple conditions

[lindsa...@gmail.com]

Hi team!

I have an experiment with different small molecule treatments, each with multiple doses, and multiple timepoints, with biological replicates of each combinatorial set of conditions (drug+dose+time). 

How should I best account for these conditions? Should I "just" concatenate the combination of conditions (e.g. drug+dose+time)? Obviously, as always, happy to share data if that's helpful :) 

Thank you!

Lindsay

[Devon Kohler]

Hi Lindsay,

This actually sounds like an interesting use case for MSstatsResponse. What sort of outputs are you looking to get from the experiment? Hit detection/IC50 estimates?

In MSstatsResponse you could combine the drug+time and look at curves over dose, but you could also model drug+concentration and look at curves over time.

Otherwise for standard MSstats you would just have to combine all three modalities and specify the comparisons with a (probably super big) contrast matrix.

I'll let Sarah comment on the feasibility of MSstatsResponse for this data.

Best,

Devon

[szve...@gmail.com]

Hi Lindsay, 

As Devon mentioned, MSstatsResponse could be useful here. MSstatsResponse works directly with the output of the 'dataProcess' function in MSstats. Here is the full vignette: https://bioconductor.org/packages//release/bioc/vignettes/MSstatsResponse/inst/doc/MSstatsResponse.html  . 

The most important step would be formatting your data using the 'MSstatsPrepareDoseResponseFit' function. Here you define the dose column (which could be dose or time) and the drug column. In your case, the drug column would be used to group the data. For example, if you're interested in analyzing drug+time over dose, you would set your drug id's to be something like "DrugA_1hr", "DrugA_3hr", etc.. This would model a curve for each protein across doses for every drug-timepoint pair in your dataset. The same approach works with drug+dose combinations. 

Once the formatting is set, MSstatsResponse supports hypothesis testing for significant drug-protein interactions, IC/EC50 estimates with confidence intervals, and visualization of dose-response curves. 

If you'd like to share some data with me, I'd be happy to write a short script that does this for you. If you have a subset of the data after running the dataProcess function that would work great. 

Best,

Sarah