doPlotFCT: Calculate fold change and t test, the plot

Description Usage Arguments Details Value Author(s) Examples


Calculate fold changes and p values from t test, and plot the results using preset FDR threshold


doPlotFCT(eset, group, grpMember, order1 = NULL, order2 = NULL,
detectSample = 0.5, snThresh = 3, ...)



an ExpressionSet object


which group from experiment design should calculation and plot be performed


optional group member within the group


optional, For a pairwise comparison the ordering of the first group of replicates


optional, For a pairwise comparison the ordering of the first group of replicates


optional number between 0 and 1 to indicate the percentage of arrays should be above snThresh to include in the t test analysis. Default = 0.5. If the probe is detected in 50% or more samples on one of the subgroup, the probe is included in the t test, otherwise, it will be excluded in the t test


optional S/N ratio threshold. Default = 3


Additional argument, currently not implemented


Group members are optional. For example, if group name is "tissue", and group members in experiment design file include "brain", "liver", "lung", "muscle". We could include c("brain", "liver") as group member for the parameter, then t test will be performed between "brain" and "liver", and "lung" "muscle" will be ignored. However, if we omit group member in the arguments, all tissue members will be used for t test. In this case, there will be 6 pairwise t test between each member of the group.

If order1 and order2 are specified then a paired sample t-test will be conducted between the groups, with the arrays in each group sorted according to the ordering specified. For example, if order1 is c(1,3,2) and order2 is c(1,2,3), then the sample pairing is a1-b1, a3-b2, a2-b3, with a and b are subgroup 1 and subgroup 2 within the group.

The fold changes are difference between averaged subgroup1 expression vs averaged subgroup2. If paired t test is performed, the fold changes are calculated using each paired difference and take an average of paired difference.


None. But a number of plot and result files will be produced.


Y Andrew Sun


#- doPlotFCT(eset, "sampleGroup", c("liver", "muscle"))
#- For a paired t test
#- doPlotFCT(eset, "sampleGroup", c("liver", "muscle"), order1 = c(1,2,3), order2 = c(1,3,2))

ABarray documentation built on May 2, 2018, 3:42 a.m.