factors and contrast matrix

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

factors and contrast matrix

frymor
Hi,

at the moment I'm trying to understand the limma/affy Packages. I've tried
every example I could find to try and understand the ways to create design
and contrast matrices. I must admit though, that I can't figure it out. I
did a few times the estrogen files from the book "bioinformatics and
computational biology solutions using R and Bioconductor".
I know it's a difficult question to answer, but I need all the help I can
get.
1. how does one suppose to build the design matrix by knowing what one wants
to examine? this is the created matrix in the book:
             Intercept ES T48 ES:T48
low10-1.cel          1  0   0     0
low10-2.cel          1  0   0     0
high10-1.cel         1  1   0     0
high10-2.cel         1  1   0     0
low48-1.cel          1  0   1     0
low48-2.cel          1  0   1     0
high48-1.cel         1  1   1     1
high48-2.cel         1  1   1     1
attr(,"assign")
[1] 0 1 2 3
attr(,"contrasts")
attr(,"contrasts")$"factor(estrogen)"
[1] "contr.treatment"
attr(,"contrasts")$"factor(time.h)"
[1] "contr.treatment"

2. There are some design matrices where the first column is full of '1'
(intercept). What does it mean? can I do without this column?

3. How do I create the contrast matrix if I want to examine the differences
in the  gene expression between 10 hours and 48 hours?
ES - with estrogen in 10 hours
T48 - time 48 hours
ES:TS - probes with estrogen and in 48 hours

(Just by the way:
4. does everything has to do with matrices multiplications? )

the book propose this contrast matrix:

> contM <- cbind(es10=c(0,1,0,0), es48=c(0,1,0,1))
> contM
     es10 es48
[1,]    0    0
[2,]    1    1
[3,]    0    0
[4,]    0    1

5. But why? How do I suppose to know it? Is it also a multiplication of some
matrices?
6. What stands the columns or the lines for?

I'll be happy for every kind of help I can get. If someone know where I can
read about it, it would be also of great need.

THX,

Assa
--
Assa Yeroslaviz
Loetzener Str. 15
51373 Leverkusen

        [[alternative HTML version deleted]]

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html