clusterSignifFDR-class       package:goCluster       R Documentation

_C_l_a_s_s "_c_l_u_s_t_e_r_S_i_g_n_i_f_F_D_R"

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

     This class provides a selection mechanism that uses the false
     dicovery rate (FDR) to identify annotation terms which are
     significantly enriched in selected gene groups.

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

     The class provides a wrapper around the 'randomizeTree',
     'mergeAnno', and 'selectAnnoStats' function. Please read the
     corresponding documentation for further details.

_S_l_o_t_s:

     '_t_h_r_e_s_h_o_l_d': Object of class '"numeric"', this specifies the
          fraction of selected annotation elements that would also be
          identified from random data. The class will select as many
          elements as possible until this threshold is reached.

     '_r_a_n_d_o_m_s_t_a_t': Object of class '"list"', a list of p-values for a
          number of randomized datasets with the same  structure as the
          original data.

     '_r_e_p_e_a_t_s': Object of class '"numeric"', specifies the number of
          randomized datasets that will be generated to  calculate the
          false discovery rate.

     '_p_t_h_r_e_s_h_o_l_d_s': Object of class '"numeric"', these are the
          p-value-thresholds for the selection. The FDR-threshold (see
          above) is transformed into a p-value threshold by 
          determining how many elements can be selected without
          exceeding  the given threshold of false positives. Since this
          is done for  each annotation dataset this holds a vector of
          thresholds.

     Additional slots are described in the documentation of the
     'clusterSignif-class' and 'clusterModule-class'.

_E_x_t_e_n_d_s:

     Class '"clusterSignif"', directly. Class '"clusterModule"', by
     class '"clusterSignif"'.

_M_e_t_h_o_d_s:

     _c_o_n_f_i_g 'signature(object = "clusterSignifFDR")': interactive setup
          of the class. You can set the FDR threshold  here.

     _s_e_t_u_p 'signature(object = "clusterSignifFDR")': returns the
          configuration of the object as a list. This list can be used
          for the non-interactive setup of the class. 

     _s_e_t_u_p<- 'signature(object = "clusterSignifFDR")': non-interactive
          setup of the class. You need to provide a list that contains
          the necessary settings for the class. 

     _e_x_e_c_u_t_e 'signature(object = "clusterSignifFDR")': selects as many
          annotation terms as possible without exceeding  the specified
          FDR. 

     _r_e_s_e_t 'signature(object = "clusterSignifFDR")': resets the results
          of this class so that the selection process can be run again. 

     _p_r_i_n_t 'signature(object = "clusterSignifFDR")': This function
          prints some basic information about the content of this
          object. If the object has been executed, it will show the
          number of identified annotation terms. 

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

     Gunnar Wrobel, <URL: http://www.gunnarwrobel.de>.

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

     'selectAnnoStats', 'randomizeTree', 'mergeAnno',
     'goCluster-class', 'clusterSignif-class', 'clusterModule-class'

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

     ## Load a small test dataset
     data(benomylsetupsmall)

     ## Create an emty goCluster object
     test <- new("goCluster")

     ## The "benomylsetupsmall" configuration select FDR selection of
     ## annotation terms
     benomylsetupsmall$classsign

     ## But we need to select a slightly higher threshold in order
     ## to find some significant annotation in this very reduced
     ## dataset
     benomylsetupsmall$sign$threshold <- 0.2

     ## Assign the configuration to the object and directly execute it
     execute(test) <- benomylsetupsmall

     ## These are the selected annotation terms
     test@sign@selection

