Changes in version 1.3 (2018-05-23)                   

  - Fixed bug, now ensuring absolute value of correlation parameter r is
    used to compute enrichment score (credit to Michael from Berkeley).

                  Changes in version 1.1 (2016-08-06)                   

  - Now includes GO term gene sets data object (i.e. a list of character
    vectors, named by GO term, of genes associated with each term - can
    be loaded with data(GO_gene_sets) [based on Gene Ontology
    (2015-11-27) and human GO annotations file validated
    on 07/08/2016]).
  - Change in how S argument of gset is interpreted when vector of
    scores r is supplied - now instead of the scores for S being those
    at the corresponding ranks in r, they are those at corresponding
    indices. For example, in version 1.0 the scores for S=1:3 would be
    the top 3 scores in r, however now it be the those of the first 3
    elements of r. Behaviour has not changed for when S is character or
    when N is used instead of r.