Isoallfdr              package:IsoGene              R Documentation

_O_b_t_a_i_n_i_n_g _t_h_e _d_e_l_t_a _t_a_b_l_e _i_n _t_h_e _S_A_M _p_r_o_c_e_d_u_r_e

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

     The function obtains the delta table in the SAM procedure for the
     five test statistics (the global likelihood test, Williams,
     Marcus, M, and the modified M).

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

     Isoallfdr(qqstat, ddelta, stat)

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

  qqstat: output from function Isoqqstat containing the test statistics
          of permutations

  ddelta: give a list of values as cut-off to find the number of
          significant genes in the SAM procedure.  If unspecified, the
          default value is assigned using the centiles of the absolute
          difference between the observed and expected test statistics.

    stat: choose one of the five test statistics to use

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

     dtable: the delta table in the SAM procedure containing six
     columns. The first column is the cut-off value to find the number
     of significant genes, the second column is the meidan number of
     false positives, the third column is the 90% percentile number of
     false positives, the fourth column is the number of significant
     genes, the fifth column is the median FDR, and the last column is
     the 90% FDR.

_N_o_t_e:

     This function calculates the delta table in the SAM procedure for
     the five test statistics. To use the SAM procedure, the number of
     genes in the dataset is preferably larger than 500.

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

     Lin et al.

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

     'isoreg', 'Isoqqstat', 'Isoqval','IsoTestSAM', 'IsoSAMPlot'

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

       set.seed(1234)
       x <- c(rep(1,3),rep(2,3),rep(3,3))
       y1 <- matrix(rnorm(9000, 1,1),1000,9) ## 1000 genes with no trends
       y2 <- matrix(c(rnorm(3000, 1,1),rnorm(3000,2,1),rnorm(3000,3,1)),1000,9) ## 1000 genes with increasing trends
       y <- data.frame(rbind(y1, y2)) ##y needs to be a data frame
       qqstat <- Isoqqstat(x, y, fudge="pooled",niter=100)
       allfdr <- Isoallfdr(qqstat,,stat="E2")
      

