Entering edit mode
Ruppert Valentino
▴
270
@ruppert-valentino-1376
Last seen 10.2 years ago
Hello,
I have a question regarding the correct use of glm for
logistic regression. I am trying to analyse microarray data of 30
cases with 600 affymetrix probes. The 30 cases have clinical
parameters on dead/alive and drug treatment/no treatment.
I enclose subset example of how the table is for 6 affymetrix probes
on 17 cases :
probe 1
probe 2
probe 3
probe 4
probe 5
probe 6
Dead/Alive
Drug/NoDrug
5.9
6.6
6.5
6.8
5.2
7.1
0
1
8.6
7.9
7.0
7.1
6.9
7.0
1
0
4.9
5.5
4.6
4.3
4.4
4.3
1
1
4.3
5.5
4.9
5.9
6.8
4.2
0
1
7.1
5.3
6.1
6.7
7.4
7.0
0
1
9.0
10.8
4.5
4.5
5.5
5.7
0
0
3.2
2.3
2.8
2.3
5.3
2.3
0
1
3.6
2.7
2.7
2.9
3.2
2.9
1
1
6.4
4.0
2.4
5.6
2.5
3.0
0
0
5.4
2.3
2.3
3.2
2.3
2.3
1
0
3.3
3.7
3.0
3.6
3.7
4.9
0
1
5.1
4.5
3.2
3.6
5.5
3.8
0
0
My question is can I use glm to see whether dead/alive or drug/nodrug
is significant in the model in multivariate way. Meaning can I take
probes 1 to 6 and compare those as matrix against Dead/Alive and
Drug/NoDrug in multivariate analysis?
so in the above example of the data.frame is called glmana :
dead/alive <- glmana[,7]
drug/nodrug <- glmanap[,8]
summary(glm(glmana[,1:6] ~ dead/alive + drug/nodrug))
Is this correct and if so could I put all 600 probes and compare that
as a matrix against dead/alive and drug/nodrug, if not is there anyway
to determine the maximum number of probes to use in the model?
Many thanks,
Ruppert
[[alternative HTML version deleted]]