I am analysing some proteomics data from an experiment with the following experimental design:

• 3 donors, each treated with:
  • placebo - 3 technical replicates
  • drug - 3 technical replicates

    sample_id   patient_id   treatment
    1                   A                  placebo
    2                   A                  placebo
    3                   A                  placebo
    4                   B                  placebo
    5                   B                  placebo
    6                   B                  placebo
    7                   C                  placebo
    8                   C                  placebo
    9                   C                  placebo
    10                 A                   drug
    11                 A                   drug
    12                 A                   drug
    13                 B                   drug
    14                 B                   drug
    15                 B                   drug
    16                 C                   drug
    17                 C                   drug
    18                 C                   drug
    

I am interested in answering the following question, for each peptide: Is there a difference between treatment and placebo? I would like to take into account the technical replicates (i.e. patient_id; treatment combination) and the fact that in both treatments samples were coming from the same donors

I am not sure how the design.matrix should look like

Many thanks

Andrea

Probably

design <- model.matrix(~treatment + patient)

Although if the technical replicates are just re-runs of the same sample prep you might consider

design <- model.matrix(~treatment)
fit <- lmFit(<data goes here>, design, block = patient)

To fit a GLS model instead.