spotseg           package:spotSegmentation           R Documentation

_M_i_c_r_o_a_r_r_a_y _S_p_o_t _S_e_g_m_e_n_t_a_t_i_o_n

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

     Microarray spot segmentation via model-based clustering.

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

     spotseg(chan1, chan2, rowcut, colcut, R=NULL, C=NULL, 
                       threshold=100, hc=FALSE, show=FALSE)

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

   chan1: matrix of pixel intensities from the first channel.

   chan2: matrix of pixel intensities from the second channel.

  rowcut: row delimiters for the spots. Entries are the starting row
          location in the close of each spot, with the last entry being
          one pixel beyond the border of the last spot. For example,
          from the output of 'spotgrid'.

  colcut: column delimiters for the spots. Entries are the starting
          column location in the close of each spot, with the last
          entry being one pixel beyond the border of the last spot. For
          example, from the output of 'spotgrid'.

       R: rows over which the spots are to be segmented. The default is
          to segment spots in all rows.

       C: columns over which the spots are to be segmented. The default
          is to segment spots in all columns.

threshold: connected components of size smaller than 'threshold' are
          ignored. Default: 'threshold=100'.

      hc: logical variable indicating whether or not EM should be
          initialized by hierarchical clustering or quantiles  in
          model-based clustering. The default is to use quantiles 'hc =
          FALSE', which is more efficient both in terms of speed and
          memory usage.

    show: logical variable indicating whether or not to display the
          segmentation of each individual spot as it is processed. The
          default is not to display the spots 'show = FALSE'.

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

     There are 'plot' and 'summary' methods that can be applied to the
     result.

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

     An array of the same dimensions as the image in which the  pixels
     are labeled according to their group within the spot area:
     1=background,2=uncertain,3=sample.

_N_o_t_e:

     The 'mclust' package is requiredfor clustering.

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

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

     Q. Li, C. Fraley, R. Bumgarner, K. Y. Yeung, and A. Raftery\
     Robust model-based segmentation of microarray images,\  Technical
     Report No.~473, Department of Statistics, University of
     Washington, January 2005.

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

     'summary.spotseg', 'plot.spotseg', 'spotgrid'

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

     data(spotSegTest)

     # columns of spotSegTest:
     #  1 intensities from the Cy3 (green) channel
     #  2 intensities from the Cy5 (red) channel

     dataTransformation <- function(x) (256*256-1-x)^2*4.71542407E-05 

     chan1 <- matrix(dataTransformation(spotSegTest[,1]), 144, 199)
     chan2 <- matrix(dataTransformation(spotSegTest[,2]), 144, 199)

     Grid <- spotgrid( chan1, chan2, rows = 4, cols = 6, show = TRUE)

     library(mclust)

     Seg <- spotseg( chan1, chan2, Grid$rowcut, Grid$colcut)

     plot(Seg)

     spotSummary <- summary(Seg)

     spot11 <- spotseg( chan1, chan2, Grid$rowcut, Grid$colcut, 
                         R = 1, C = 1, show = TRUE)

