Version 0.25:
  * corrected the malfunctioning testthat test

Version 0.24:
  * msigDB is broken, it no longer has unique identifiers in the global XML
    file. I corrected the import function to make sure identifiers are
    unique.
  * pvalEffectPlot was expecting log10(p-values) rather than p-values. This
    has been taken care of.
  * evidencePlot can now generate GSEA-style curves (with the "style"
    option)


Version 0.22:

  * added Cohen's d estimate and t statistic to tmodLimmaTopTable()
  * rewrote evidencePlot: it is now much more flexible
