VERSION 1.9.6
-------------------------

  o BEHAVIOUR CHANGE: testEnrichmentSEA reports nQ as the size of the query
    and nD as the size of the database. For a numeric query the two were the
    other way round (nQ was the set, nD the CpGs with a value).
  o testEnrichment warns when no query probe is in the universe, instead of
    returning all-zero overlaps silently.
  o Databases dropped while loading are named in a warning: a requested group
    with no version for the query's platform, an unknown group, an empty set.
  o Vignettes: no sesameDataCache("HM450.address"); sesameData 1.31.2 lists
    that record with no ExperimentHub id yet, which stopped the devel build.

VERSION 1.9.5
-------------------------

  o testEnrichment: no "Unknown or uninitialised column" warnings on
    sequencing results; the estimate's 1024 / -1022 sentinels are documented.
  o Vignettes: an install chunk for the Suggests the examples load; the
    Manhattan and feature-aggregation chunks cache their own data; MM285
    examples use the MM285 TFBS database; the HM450 Zenodo example uses an
    HM450 query; the sequencing quick start runs on the shipped single cell;
    the bedToCg example is self-contained; permutation chunks are seeded.
  o Manual: loadDBs, testEnrichmentSpearman and df_master are internal; no
    developer notes or local paths in examples.

VERSION 1.9.4
-------------------------

  o bedToCg: the bedtools/yame pipeline reached sh -c unquoted, so sh ran a
    bare `yame` and the output file was empty. It is now quoted.
  o testEnrichment2: the odds ratio is computed in double precision. Integer
    counts from YAME overflowed on genome-scale tables and gave NA estimates.
  o testEnrichment2: an unmeetable min_overlap returns the empty result
    without a second warning.
  o KYCG_plotEnrichAll: geom_segment(linewidth=), not the deprecated size=.

VERSION 1.9.3
-------------------------

  o testEnrichmentSEA: the two tails now share one pooled permutation
    null, so the p-value no longer depends on the sign of the query
    (zhou-lab/knowYourCG#6)
  o testEnrichmentSEA: the sign of estimate is flipped to match GSEA.
    Positive now means the set is enriched at large values, negative at
    small values. Results from earlier versions have the opposite sign.
    KYCG_plotSetEnrichment now walks from the largest value down.

VERSION 0.99.0
-------------------------

  o First submission of knowYourCG package to Bioconductor
