### Version 1.2.1

# Bug fix

1 - ghap.haplotyping. If a block had no HapAllele the function would stop.
2 - ghap.blmm. If IDs had a pattern matching with the random column name, the function would stop.
3 - ghap.blmm. Weights are not used if ordinal=T, but the function would stop anyway in the absence of weights.
4 - ghap.blmm. Heritability values were computed based on iterations rather than posterior estimates of variance components.
5 - ghap.kinship. Malfunctioning of batching.
6 - ghap.kinship and ghap.blup. The scaling factor q (inverse sum of the variances in the columns of the HapGenotype matrix) was not weighted by HapAllele specific weights.
7 - vignette. The static access to the vignette was broken.
8 - ghap.assoc. HapBlock test parameterization has been changed to avoid singular H'H matrices.

# Inclusion of new utilities

1 - ghap.mme. Given known variance components, this function speeds up computation of fixed and random effects by simply solving Henderson's mixed model equations.
2 - ghap.blup. Now outputs the mean genotype for each HapAllele (meaningful for predictions with ghap.profile).
3 - ghap.profile. Now allows for centering and scaling of genotypes using arbitrary constants.
4 - ghap.lmm. This function uses L-BGFS-B minimization to obtain maximum likelihood estimates of mixed model parameters.
5 - The package has now a startup message giving instructions to access the vignette, as well as information regarding citation.
6 - ghap.simpheno. Extended capabilities to scale the total phenotypic variance and fine tune variance explained by specific haplotypes.

### Version 1.2.2

# Inclusion of new utilities

1 - The Matrix package is now used to handle big and sparse matrices
2 - ghap.blmm, ghap.mme and ghap.lmm have been replaced by a single ghap.lmm function Bayesian models will be re-implemented in the future.
3 - The lme4 package is now used to fit generalized linear models.
