Standards with MSstats

[Christian Potts]

Hello MSstats team,

I was with you during May Institute last month and would like clarification on what you shared about using standards with MSstats. I got the impression that the globalstandards normalization was the most robust option for normalization. We're conducting a DIA experiment on some mouse tissue samples, and I'm not sure whether I should add standards. In the lectures, it was shared that there are no commercial standards for experiments on the proteome scale (1000s of proteins), and "hybrid proteome" or in-house mixes are used. What do those refer to? Could you please share some wisdom to help me understand?

Thank you for the MSstats help! 
Christian Potts 

[Anthony Wu]

Hey Christian,

When we mention "hybrid proteome" or in-house mixes, that should be referring to how one could spike in the proteome of another species (e.g. spiking in a human digest in your mouse sample) and normalize assuming the peptides of the spiked-in species is constant.  One thing about reference standards is that they can normalize technical effects that occur after spiking in, but any effects occurring before spike-in can't be adjusted for.   

Given you're doing a DIA experiment, you could normalize by equalizing the medians of the log intensities of the runs.  Median normalization would account for all technical effects that occur at every stage of prepping the sample to running the mass spec.  But the downside is that it assumes that most proteins are not changing across your conditions of interest.

Hope this helps,

Tony