[BioC] maanova rss for null model

[James MacDonald]

Hi Brendan,

maanova is not part of Bioconductor, so you are looking in the wrong
place. If you want help with maanova, you should probably contact Hao
Wu, who is the guy who wrote all the functions.

hao at jax.org

Alternatively, you might be able to find what you are looking for in
the maanova user's manual.


Jim



James W. MacDonald
UMCCC Microarray Core Facility
1500 E. Medical Center Drive
7410 CCGC
Ann Arbor MI 48109
734-647-5623

>>> "Brendan Heavey" <bmheavey at buffalo.edu> 06/30/03 01:18PM >>>

Hello-

I'm trying to recreate the calculation maanova goes through on my data.
 In
the paper, it shows

    (rss0-rss1)/(df0-df1)]
F1= ----------------------
          (rss1/df1)

I have 6 biological replicates for each gene, and they're all spotted
3
times on each slide, and there is a dye flip (Design:
SampleID-1221-1221-1221-1221-1221-1221)

Maanova calculates a F1 score for each gene in the data.  I believe I
know
how to calculate rss1 and df1 for each gene by the following:

1). Calculate Overall Mean for that gene
2). Subtract signal value from each spot signal across all 3 animals
from #1
3). Square #2
4). Sum #3

I'm running into trouble trying to understand how to calculate rss0. 
I
guess I'm confused about what the null model really represents...I
thought I
would just find the overall mean from all spots/animals, subtract each
signal from that mean then sum the squares.  This doesn't appear to be
working.

Thanks ahead of time for any help you can provide...i greatly
appreciate it

_______________________________________________
Bioconductor mailing list
Bioconductor at stat.math.ethz.ch 
https://www.stat.math.ethz.ch/mailman/listinfo/bioconductor