limma vs groupComparison

[s.malek...@sheffield.ac.uk]

I have a  question regarding SRM significance analysis using MSStas, 

I have SRM data set including case and control.

I have used two approach:

1. I used MSStas extracting  protein Intensity  then used limma to find  significant protein

2. In Second  this approach I have used :

 resultOneComparison<-groupComparison(contrast.matrix=comparison,data=quantData)

Why do I get different result? Could you please explain what method has been used in MSStats for group comparison.

Regards, Afsaneh 

[Tsung-Heng Tsai]

Here are some details of MSstats including the modeling and parameter estimation:

http://msstats.org/wp-content/uploads/2016/03/UShupo2016_short_course_MSstats.pdf

A difference between limma and MSstats is how the variance estimate is obtained.

[Olga Vitek]

To add to Tsung-Heng's comments: even though in your example MSstats and limma use the same input, these are different models (in particular with respect to variance estimation) and therefore produce different results. If you would like to know all the details, please join us at the May Institute https://computationalproteomics.ccis.northeastern.edu/ where we will discuss this extensively.

HTH

Olga