emfit                package:EBarrays                R Documentation

_I_m_p_l_e_m_e_n_t_s _E_M _a_l_g_o_r_i_t_h_m _f_o_r _g_e_n_e _e_x_p_r_e_s_s_i_o_n _m_i_x_t_u_r_e _m_o_d_e_l

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

     Implements the EM algorithm for gene expression mixture model

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

     emfit(data,
           family,
           hypotheses,
           theta.init = NULL, p.init = NULL,
           num.iter = 20,
           verbose = getOption("verbose"),
           trace = TRUE,
           optim.control = list())

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

    data: a matrix or an an object of class '"exprSet"' or
          '"ExpressionSet"'

  family: an object of class ``ebarraysFamily'' or a characterstring
          which can be coerced to one. Currently, only the characters
          "GG" and "LNN" are valid. Other families can be supplied by
          constructing them explicitly. 

hypotheses: an object of class ``ebarraysPatterns'' representing the
          hypotheses of interest. Such patterns can be generated by the
          function 'ebPatterns' 

theta.init: initial values of theta. Should be appropriate for the
          corresponding 'family'. If omitted, defaults are calculated
          based on the 'family'. 

  p.init: initial values of mixture probabilities. If omitted, all
          hypotheses are initially taken as equally likely. 

num.iter: number of EM iterations

 verbose: logical or numeric (0,1,2) indicating desired level of
          information printed for the user

   trace: whether the result should have slots 'thetaTrace' and
          'probTrace', matrices giving values for all iterations 

optim.control: list passed unchanged to 'optim' for finer control 

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

     an object of formal class ``ebarraysEmFit'' with the following
     slots

'family': the 'family' passed to 'emfit' 

'hypotheses': the 'hypotheses' passed to 'emfit' 

'thetaEst': numeric vector holding the parameter estimates 

'probEst': numeric vector holding estimates of the mixture
          probabilities 

'thetaTrace': optional matrix holding intermediate values of 'thetaEst' 

'probTrace': optional matrix holding intermediate values of 'probEst' 


     The 'show()' method can be used to summarize the object.  It can
     also be used to generate posterior probabilities using 'postprob'

_N_o_t_e:

     'emfit' is a generic function dispatched on the arguments 'data',
     'family' and 'hypotheses'.

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

     Christina Kendziorski, Michael Newton and Deepayan Sarkar

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

     UW Madison Biostatistics and Medical Informatics Technical Report
     166, ``On parametric empirical Bayes methods for comparing
     multiple groups using replicated gene expression profiles,'' by
     C.M. Kendziorski, M.A. Newton, H. Lan, and M.N. Gould.  See <URL:
     http://www.biostat.wisc.edu/~kendzior/>

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

     'ebPatterns', 'ebarraysFamily-class'

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

     data(sample.exprSet.1) ## from Biobase
     patterns <- ebPatterns(c("1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1",
                              "1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2"))
     gg.fit <- emfit(data = sample.exprSet.1, family = "GG", hypotheses = patterns,
                     verbose = TRUE)
     ##                theta.init = c(10, 1, 1), p.init = c(.99, .01))
     show(gg.fit)

