[BioC] limma - dupcor.series

[Gordon Smyth]

At 05:01 AM 19/04/2004, Helen Cattan wrote:
>Hi,
>
>Thanks for the reply. I've typed in what you suggested for more
>information. Done with and without background correction. I don't know
>what you are looking for - but hope this helps.

The summary output shows that all your data is zero or NA after background 
correction, so you can hardly expect to do any linear modelling with it. 
Even without background correction, half your data is missing. You could 
update to the current version of limma from 
http://bioinf.wehi.edu.au/limma, which may be more robust in the presence 
of NAs, but I think you basically need to go back to your original data 
files and ask why there is so much missing data. Foreground and background 
intensities from Genepix should never be missing.

>After using RG=backgroundCorrect(RG, method="minimum")
>
>summary(MAmin2$M)
>  QCmouse..CD4.depl.day.7..260803...590.530..liver.
>  Min.   :0
>  1st Qu.:0
>  Median :0
>  Mean   :0
>  3rd Qu.:0
>  Max.   :0
>
>  QCmouse..CD4.depl.day.8..10903.II....590.550..liver.4
>  Min.   :0
>  1st Qu.:0
>  Median :0
>  Mean   :0
>  3rd Qu.:0
>  Max.   :0
>
>  QCmouse.liver..CD4.depl..1.6.ug..day.8..670.530..240703
>  Min.   :0
>  1st Qu.:0
>  Median :0
>  Mean   :0
>  3rd Qu.:0
>  Max.   :0

All the above columns are entirely zero!

>
>  QCmouse.liver.CD4.depl.day.8..10903.I....590.540..
>  Min.   :   NA
>  1st Qu.:   NA
>  Median :   NA
>  Mean   :  NaN
>  3rd Qu.:   NA
>  Max.   :   NA
>  NA's   :31200

This column is entirely missing!

>summary(MAmin2$weights)
>QCmouse..CD4.depl.day.7..260803...590.530..liver.
>  Min.   :0.0000
>  1st Qu.:1.0000
>  Median :1.0000
>  Mean   :0.9873
>  3rd Qu.:1.0000
>  Max.   :1.0000
>  QCmouse..CD4.depl.day.8..10903.II....590.550..liver.4
>  Min.   :0.000
>  1st Qu.:1.000
>  Median :1.000
>  Mean   :0.986
>  3rd Qu.:1.000
>  Max.   :1.000
>  QCmouse.liver..CD4.depl..1.6.ug..day.8..670.530..240703
>  Min.   :0.0000
>  1st Qu.:1.0000
>  Median :1.0000
>  Mean   :0.9762
>  3rd Qu.:1.0000
>  Max.   :1.0000
>  QCmouse.liver.CD4.depl.day.8..10903.I....590.540..
>  Min.   :0.0000
>  1st Qu.:1.0000
>  Median :1.0000
>  Mean   :0.9878
>  3rd Qu.:1.0000
>  Max.   :1.0000
> >
> > summary(cor$cor.genes)
>Length  Class   Mode
>      0   NULL   NULL
> >
>And not using background correct
>
> > summary(MA1$M)
>  QCmouse..CD4.depl.day.7..260803...590.530..liver.
>  Min.   :-5.188e+00
>  1st Qu.:-2.813e-01
>  Median :-1.425e-02
>  Mean   : 2.424e-02
>  3rd Qu.: 2.949e-01
>  Max.   : 5.752e+00
>  NA's   : 1.326e+04
>  QCmouse..CD4.depl.day.8..10903.II....590.550..liver.4
>  Min.   :  -5.25729
>  1st Qu.:  -0.24996
>  Median :  -0.03117
>  Mean   :   0.09563
>  3rd Qu.:   0.35874
>  Max.   :   6.01569
>  NA's   :8209.00000
>  QCmouse.liver..CD4.depl..1.6.ug..day.8..670.530..240703
>  Min.   :-7.200e+00
>  1st Qu.:-2.703e-01
>  Median : 4.189e-03
>  Mean   : 3.571e-02
>  3rd Qu.: 3.068e-01
>  Max.   : 1.293e+01
>  NA's   : 1.611e+04
>  QCmouse.liver.CD4.depl.day.8..10903.I....590.540..
>  Min.   :-7.880e+00
>  1st Qu.:-2.729e-01
>  Median :-5.434e-04
>  Mean   : 8.506e-02
>  3rd Qu.: 3.162e-01
>  Max.   : 1.025e+01
>  NA's   : 1.808e+04

Have of your values are missing! If you have indeed not background 
corrected then this shows that half of your foreground intensities are missing!

Gordon

> > summary(MA1$weights)
>  QCmouse..CD4.depl.day.7..260803...590.530..liver.
>  Min.   :0.0000
>  1st Qu.:1.0000
>  Median :1.0000
>  Mean   :0.9873
>  3rd Qu.:1.0000
>  Max.   :1.0000
>  QCmouse..CD4.depl.day.8..10903.II....590.550..liver.4
>  Min.   :0.000
>  1st Qu.:1.000
>  Median :1.000
>  Mean   :0.986
>  3rd Qu.:1.000
>  Max.   :1.000
>  QCmouse.liver..CD4.depl..1.6.ug..day.8..670.530..240703
>  Min.   :0.0000
>  1st Qu.:1.0000
>  Median :1.0000
>  Mean   :0.9762
>  3rd Qu.:1.0000
>  Max.   :1.0000
>  QCmouse.liver.CD4.depl.day.8..10903.I....590.540..
>  Min.   :0.0000
>  1st Qu.:1.0000
>  Median :1.0000
>  Mean   :0.9878
>  3rd Qu.:1.0000
>  Max.   :1.0000
> >
> > summary(cor$cor.genes)
>Length  Class   Mode
>      0   NULL   NULL
>
>
>
>Thanks,
>Helen