Changes in 1.3.1:
* added CITATION file with the reference to the published article
* fixed typos in the vignette

Changes in 1.3:
* added new metrics: NDCG10, NDCG20, NDCG50, NDCG100. This changes the accuracies save file columns count
* now accepting fractional genotypes in Noisy Data Set objects (if explicitly allowed by the user)
* added new function phenoregressor.BGLR.multikinships to implement Bayesian GBLUP using more than one kinship matrices
* added option to set the accuracy file name (before it defaulted to "accuracy.csv")
* expanded GROAN.run interface to report the list of tested regressors

Changes in 1.2.1:
* added dependency to rmarkdown to comply to new knitr policy

Changes in 1.2:
* ridge regression BLUP now supports both SNP-BLUP and G-BLUP
* plotResult function now has the option to select what to use as labels
  for the x axis
* typos in the documentation

Changes in 1.1:
* GROAN.run now supports test on datasets different from the one used
  for training. Its interface, GROAN.Result object and plotResult function
  were changed accordingly
* a second pea dataset, GROAN.AI, was included, and the old dataset was
  refactored as GROAN.KI
* old dataset notation (GROAN.pea.SNPs, GROAN.pea.yield,
  GROAN.pea.kinship) was maintained but marked as deprecated in
  documentation
* added or improved print and summary function for all GROAN objects