[BioC] Comparing multilevel factors

Md.Mamunur Rashid mamunur.rashid at kcl.ac.uk
Tue Dec 22 14:50:34 CET 2009


Dear List,

I have Micro array data from illumina Human HT12  chip. The samples are from 3 major groups-
(e.g. Control (C) , Symptomatic (S) , Patient (P) ). Now using linear model I managed to find out
some differentially expressed genes.

Thanks to all of you again for your prompt response and patient support.

Now the samples can be be classified with few other factors
	like Gender( 'G') , Hypertension ( H ), Drinking ( D ).
So if the samples are labeled using these factors as well, it looks like following

CMHD - C(Control), M (Male) , H (Has Hypertension) , D ( Drinks)

SFNN - s(Symptomatic) , F (Female) , N - (No Hypertension), N (Does not Drink)

So the samples are basically-
sample<- c("CMHD","SFND","CFNN",......................)

The primary groups i.e. Control, Symptomatic and Patient are designed in such a way that
all the group have proportionate number of Gender, Smoker/ Non-smoker, Hypertension/ No- Hype.
This was done to eliminate amount of bias.

A comparison between three primary groups can be achieved using limma.

designMatrix<-  C 	S	P
		 1	0	0
		 0	1	0
		 0	1	0
		 0	0	1
		 ...
		 ...

If I want to include other factor (e.g. Age, smoking habit ) in the comparison (i.e want to see the
effect of this factors) in the comparison between two Main Groups, how do I do it.

Thanks in advance.
And Wish you all a Marry Christmas and Happy New Year.....!!!!

regards,
Md.Mamunur Rashid



More information about the Bioconductor mailing list