[BioC] Navigating data in the cellHTS2 object

Ian Sudbery ims at sanger.ac.uk
Wed Jun 24 13:23:15 CEST 2009


Hi,

    I'm just trying to migrate from cellHTS1 to cellHTS2 and I wonder if 
somebody could answer me some questions around the new structure of the 
object.

If x is a cellHTS2 object created with:

x <- readPlateList("platelist2.txt", name = expName, path = dataPath)

from a platelist file that look a bit like:

Filename    Plate    Replicate    Channel    Batch
1a1.before    1    1    1    1
1a2.before    1    2    1    1
1b1.before    2    1    1    1
1b2.before    2    2    1    1
2a1.before    3    1    1    1
2a2.before    3    2    1    1
2b1.before    4    1    1    2
2b2.before    4    2    1    2

where the different batches represent plates processed by different 
experimenters. The complete set have 18 plates, with 2 replicates and 2 
channels. The plates are divided into 6 batches.

and configured with a plate configuration file like:

Wells:    96   
Plates:    18   
Plate    Well    Content
*    *    sample
*    A01    neg1
*    B01    neg1
*    C01    siCasp8
*    D01    siCasp8
*    E01    siCasp8
*    F01    siBid
*    G01    siBid
*    H01    siBid
*    A12    empty
*    B12    empty
*    C12    nt
*    D12    nt
*    E12    siKif11
*    F12    siKif11
*    G12    neg2
*    H12    neg2

channels are normalised with
xn<-summarizeChannels(x,fun = function (r1,r2) r2/r1)

and plate normalised with

xnp <- normalizePlates(xn, scale = "multiplicative", log = TRUE, method 
= "median", varianceAdjust = "none")

How would I perform tasks such as

1)producing a mean/sd for each type of control well for each plate (with 
separate replicates of the same plate being counted separately), and 
each batch (to look for differences between batches).

2) Plot the correlation between the two replicates for each batch to 
find if one batch is more variable than any other.

3) Produce box plots for the sample values in each plate (again plates 
in different replicates treated as different plates) so that different 
normalization methods can be compared.

4)Average over the values for each well in every plate (i.e. all of the 
A01s or all of the B05s), output as a 8x12 matrix so that a plot can be 
produced to look for edge effects in the plates.

I'm having a little trouble getting to grips with the more complex 
structure of cellHTS2 as compared to cellHTS1.

Cheers,


Ian Sudbery
------------

Wellcome Trust Sanger Institute.



-- 
 The Wellcome Trust Sanger Institute is operated by Genome Research 
 Limited, a charity registered in England with number 1021457 and a 
 company registered in England with number 2742969, whose registered 
 office is 215 Euston Road, London, NW1 2BE.



More information about the Bioconductor mailing list