clusterGenome              package:aCGH              R Documentation

_c_l_u_s_t_e_r_i_n_g _a_n_d _h_e_a_t_m_a_p

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

     This function clusters samples and shows their heatmap

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

     clusterGenome(aCGH.obj,
                        response = as.factor(rep("All", ncol(aCGH.obj))),
                        chrominfo = human.chrom.info.Jul03, cutoff=1,
                        lowCol = "red", highCol = "green", midCol = "black",
                        ncolors = 50, byclass = FALSE, showaber = FALSE,
                        amplif = 1, homdel = -0.75,
                        samplenames = sample.names(aCGH.obj),
                        vecchrom = 1:23, titles = "Image Plot",
                        methodS = "ward", dendPlot = TRUE, imp = TRUE,
                        categoricalPheno = TRUE)

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

aCGH.obj: object of class 'aCGH' here

response: phenotype of interest. defaults to the same phenotype
          assigned to all samples

chrominfo: a chromosomal information associated with the mapping of the
          data

  cutoff: maximum absolute value. all the values are floored to
          +/-cutoff depending on whether they are positive of negative.
          defaults to 1

 ncolors: number of colors in the grid. input to 'maPalette'. defaults
          to 50

  lowCol: color for the low (negative) values. input to 'maPalette'.
          defaults to "red"

 highCol: color for the high (positive) values. input to 'maPalette'.
          defaults to "green"

  midCol: color for the values close to 0. input to 'maPalette'.
          defaults to "black"

 byclass: logical indicating whether samples should be clustered within
          each level of the phenotype or overall. defaults to F

showaber: logical indicating whether high level amplifications and
          homozygous deletions should be indicated on the plot.
          defaults to F

  amplif: positive value that all observations equal or exceeding it 
          are marked by yellow dots indicating high-level changes.
          defaults to 1

  homdel: negative value that all observations equal or below it  are
          marked by light blue dots indicating homozygous deletions.
          defaults to -0.75

samplenames: sample names

vecchrom: vector of chromosomal indeces to use for clustering and to
          display. defaults to 1:23

  titles: plot title. defaults to "Image Plots"

 methodS: clustering method to cluster samples. defaults to "ward"

dendPlot: logical indicating whether dendogram needs to be drawn.
          defaults to T.

     imp: logical indicating whether imputed or original values should
          be used. defaults to T, i.e. imputed.

categoricalPheno: logical indicating whether phenotype is categorical.
          Continious phenotypes are treated as "no groups" except that
          their values are dispalyed.defaults to TRUE. 

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

     This functions is a more flexible version of the 'heatmap'. It can
     cluster within levels of categorical phenotype as well as all of
     the samples while displaying phenotype levels in different colors.
     It also uses any combination of chromosomes that is requested and
     clusters samples based on these chromosomes only. It draws the
     chromosomal boundaries and displays high level changes and
     homozygous deletions. If phenotype if not categical, its values
     may still be displayed but groups are not formed and byclass = F.
     Image plot has the samples reordered according to clustering
     order.

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

     'aCGH' 'heatmap'

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

      
     data(colorectal)

     #cluster all samples using imputed data on all chromosomes (autosomes and X):

     clusterGenome(colorectal)

     #cluster samples within sex groups based on 3 chromosomes individually. 
     #use non-imputed data and  do not show dendogram. Indicate amplifications and 
     #homozygous deletions.

     clusterGenome(colorectal, response = phenotype(colorectal)$sex,
                        byclass = TRUE, showaber = TRUE, vecchrom = c(4,8,9),
                        dendPlot = FALSE, imp = FALSE)

     #cluster samples based on each chromosome individualy and display age. Show
     #gains in red and losses in green. Show aberrations and use values < -1
     #to identify homozgous deletions. Do not show dendogram.

     pdf("plotimages.pdf", width = 11, height = 8.5)
     for (i in 1:23)
         clusterGenome(colorectal,
                            response = phenotype(colorectal)$age,
                            chrominfo = human.chrom.info.Jul03,
                            cutoff = 1, ncolors = 50, lowCol="green",
                            highCol="red", midCol="black", byclass = FALSE,
                            showaber = TRUE, homdel = -1, vecchrom = i,
                            titles = "Image Plot", methodS = "ward",
                            dendPlot = FALSE, categoricalPheno = FALSE)
     dev.off()

