mmgmos                 package:puma                 R Documentation

_M_u_l_t_i-_c_h_i_p _m_o_d_i_f_i_e_d _g_a_m_m_a _M_o_d_e_l _f_o_r _O_l_i_g_o_n_u_c_l_e_o_t_i_d_e _S_i_g_n_a_l

_D_e_s_c_r_i_p_t_i_o_n:

     This function converts an object of class 'AffyBatch' into an
     object of class  'exprReslt' using the Multi-chip modified gamma
     Model for Oligonucleotide Signal  (multi-mgMOS). This function
     obtains confidence of measures, standard deviation and 5,  25, 50,
     75 and 95 percentiles, as well as the estimated expression levels.

_U_s_a_g_e:

     mmgmos(
             object
     ,       background=FALSE
     ,       replaceZeroIntensities=TRUE
     ,       gsnorm=c("median", "none", "mean", "meanlog")
     ,       savepar=FALSE
     ,       eps=1.0e-6
     ,       orig.phis = FALSE
     ,       addConstant = 0
     )

_A_r_g_u_m_e_n_t_s:

  object: an object of 'AffyBatch'

background: Logical value. If 'TRUE', perform background correction
          before applying mmgmos.

replaceZeroIntensities: Logical value. If 'TRUE', replace 0 intensities
          with 1 before applying mmgmos.

  gsnorm: character. specifying the algorithm of global scaling
          normalisation.

 savepar: Logical value. If 'TRUE' the estimated parameters of the
          model are saved in file par_mmgmos.txt and phi_mmgmos.txt. 

     eps: Optimisation termination criteria.

orig.phis: Logical value. If 'TRUE', use phi values created from hgu95a
          array.

addConstant: numeric. This is an experimental feature and should not
          generally be changed from the default value. 

_D_e_t_a_i_l_s:

     The obtained expression measures are in log base 2 scale.

     The algorithms of global scaling normalisation can be one of
     "median", "none", "mean", "meanlog". "mean" and "meanlog" are
     mean-centered normalisation on raw scale and log scale
     respectively, and "median"  is median-centered normalisation.
     "none" will result in no global scaling normalisation being
     applied.

     There are 2*n+2 columns in file par_mmgmos.txt, n is the number of
     chips. The first n columns are 'alpha' values for n chips, the
     next n columns are 'a' values for n  chips, column 2*n+1 is 'c'
     values and the final column is values for 'd'. The file
     phi_mmgmos.txt keeps the final optimal value of 'phi'.

_V_a_l_u_e:

     An object of class 'exprReslt'.

_A_u_t_h_o_r(_s):

     Xuejun Liu, Magnus Rattray, Marta Milo, Neil D. Lawrence

_R_e_f_e_r_e_n_c_e_s:

     Liu,X., Milo,M., Lawrence,N.D. and Rattray,M. (2005) A tractable
     probabilistic model for  Affymetrix probe-level analysis across
     multiple chips, Bioinformatics 21: 3637-3644.

     Milo,M., Niranjan,M., Holley,M.C., Rattray,M. and Lawrence,N.D.
     (2004) A probabilistic approach for summarising  oligonucleotide
     gene expression data, technical report available upon request.

     Milo,M., Fazeli,A., Niranjan,M. and Lawrence,N.D. (2003) A
     probabilistic model for the extractioin of expression levels from
     oligonucleotide arrays, Biochemical Society Transactions, 31:
     1510-1512.

     Peter Spellucci. DONLP2 code and accompanying documentation.
     Electronically available via  http://plato.la.asu.edu/donlp2.html

_S_e_e _A_l_s_o:

     Related class 'exprReslt-class' and related method 'mgmos'

_E_x_a_m_p_l_e_s:

     ## Code commented out to speed up checks
     ## load example data from package affydata
     # if (require(affydata)) data(Dilution)

     ## use method mmgMOS to calculate the expression levels and related confidence 
     ## of the measures for the example data
     #eset<-mmgmos(Dilution)

