This is a test version of Biostars. For the public version, visit https://www.biostars.org.
What is the right design matrix in a Microarray data analysis?

I have Microarray data with 3 sample each in uninfected_48hrs, infected_48hrs, uninfected_60hrs and infected_60hrs.

#3 sample infected_48h; 3 sample Uninfected_48h; 3 sample infected_60h &  sample Uninfected_60h

#First design matrix looks like this:

                           infected_48h  infected_60h  Uninfected_48h  Uninfected_60h
    Uninfected_48h.1            0            0              1              0
    Uninfected_48h.2            0            0              1              0
    Uninfected_48h.3            0            0              1              0
    infected_48h.1              1            0              0              0
    infected_48h.2              1            0              0              0
    infected_48h.3              1            0              0              0
    uninfected_60h.1            0            0              0              1
    uninfected_60h.2            0            0              0              1
    uninfected_60h.3            0            0              0              1
    infected_60h.1              0            1              0              0
    infected_60h.2              0            1              0              0
    infected_60h.3              0            1              0              0
    attr(,"assign")
    [1] 1 1 1 1
    attr(,"contrasts")
    attr(,"contrasts")$sample.levels
    [1] "contr.treatment"

  # Second design matrix
                        Uninfected_48h  infected_48h  infected_60h  Uninfected_60h
   Uninfected_48h.1              1            0            0              0
   Uninfected_48h.2              1            0            0              0
   Uninfected_48h.3              1            0            0              0
   infected_48h.1                1            1            0              0
   infected_48h.2                1            1            0              0
   infected_48h.3                1            1            0              0
   uninfected_60h.1              1            0            0              1
   uninfected_60h.2              1            0            0              1
   uninfected_60h.3              1            0            0              1
   infected_60h.1                1            0            1              0
   infected_60h.2                1            0            1              0
   infected_60h.3                1            0            1              0
   attr(,"assign")
   [1] 0 1 1 1
   attr(,"contrasts")
   attr(,"contrasts")$SKN_group_int
   [1] "contr.treatment" 

  # Which design matrix is prefer in this  scenario first design matrix or second design matrix?

Which design matrix should I used in this case and why? I am new to this design matrix concept. Please do comment and suggest thank you.

microarry designmatrix r
• 1,620 views • link
updated by JC JC • written by Paul Paodumai • Paul Paodumai

Something is not right, here. Please show all prior code, and paste samples of your metadata, if possible. Thank you.

• 0 views • link
written by Kevin Blighe Kevin Blighe

I was curious on which design matrix should I used/based on if I have 12 samples (3 infected_48hr, 3 uninfected, 3 infected_60hr and 3 uninfected_60hr). So I didn't paste the prior code. I think I should used the first design matrix.

• 0 views • link
written by Paul Paodumai • Paul Paodumai

Sure, and yes, you can now conduct individual comparisons via makeContrasts() and contrasts.fit()

• 0 views • link
written by Kevin Blighe Kevin Blighe

1 answer

in your first matrix, you have all cases separated, then you can compare directly [infected vs uninfected] or [48h vs 60h], or [infected_48h vs uninfected 48h].

In the second case, you are using your uninfected_48h as a base point, so your comparisons depend on that.

So, it depends on what are you asking your data to decide the best design matrix.

• 0 views • link
written by JC JC

Log in to answer this question.

More posts like this