# T-test for differentially expressed proteins in R?

I hope I am asking to the right place. I am currently analyzing some Mass Spectometry data of 2 mutant phenotypes and the WT with 5 biological replicates each. I have normalized the abundance values of each peptide (and its respective protein) across the samples and now I want to perform a test to identify signifficantly differentially expressed proteins. My thought was to perform standard 2-sample t-tests, comparing the measured abundances for each peptide or protein across the conditions of interest. But being a biologist, I am unsure about the assumptions I am making with these tests and how to perform them using R.

Here is an example of my data:

    Normaliz.Abundance_Sample1 Normaliz.Abundance_Sample2
Peptide A   3165565416      757986204
Peptide B   1765535403      2457586199


with the data frame having 15 columns (3x5 experiment) for around 8000 identified proteins.

Really appreciate any kind of help. R.

• You say you have 15 columns (observations) and two rows (variables), so what is the 3x5 experiment? Is there a table? T-test is used to compare the mean value of each sample, it assumes, among other things, normal distribution of the differences on the two samples, if that is not satisfied you can use a non-parametric equivalent t-test. – user2974951 Jan 9 at 7:23
• Sorry it was a lab expression (3 samples of 5 replicates each). I performed a t-test with the command dataWt_mut$pvalue<- apply(dataWt_mut, 1, function(x) t.test(x[1:5], x[6:10])$p.value) Can I get the quantified difference (something as the logFC in gene expression) from a t-test? Thanks! Edit: Format – Rina Jan 9 at 14:18
• If you have replicates (or repeated measures as they are called in statistics) then you will need to account for this in your model and a t test won't do it. A generalized linear model could do that. Also you will need to adjust the p-values as you are doing multiple comparisons. – user2974951 Jan 10 at 12:07