expression bar plot is a nice one. lets say your main data from cuffdiff is "cuff_data" and your filtered table is "my_significants_filtered". A warning: if you have a couple of hundred significantly differentially expressed genes, your graph will be too crowded. With using a better threshold for p-value, subsetting your data may results in better graphs. you can draw expr. bar plot with a command like this;

ids_to_be_graphed <- my_significants_filtered$geneids
my_genes_to_be_graphed <- getGenes(cuff_data, ids_to_be_graphed)
exp_b_plot <- expressionBarplot(my_genes_to_be_graphed)
exp_b_plot `

I dont really use cuff for rnaseq, XLOC always gives a hard time. In most cases, when you first draw the graph, you dont see the graph itself but you see gene names that are labeling the x-axis. increasing the width and height will let you see your graph.