Changes from Version 2.1.2 to 2.2.0 [September 06 2023]

* Major modifications:
** reimplementation of the compute.fstats function to improve memory usage and increase speed (> 1 order of magnitude).

* Adding two new functions:
** randomallele.pca to perform PCA on pooldata (or countdata) objects after randomly sampling one read or count per population and SNP to make the representation unaffected by differences in pool (pop) sample size and/or read coverage (or Pool-Seq data).
** pooldata2diyabc to convert a pooldata object into DIYABC genotype data file

* Other modifications
** adding computation of pairwise-population divergence (1-Q2) in compute.fstats
** adding option return.F2.blockjackknife.samples in compute.fstats to export estimates of all pairwise F2 for each block-jackknife blocks. The obtained array is compatible with functions from the admixtools2 package
** adding option output.pairwise.fst and output.pairwise.div to the compute.fstats function
** reimplementation of the compute.F4ratio function to estimate F4 ratio from F2 and F2 block-jackknife samples (rather than F4)
** fixing minor bug in vcf2pooldata to allow proper parsing of vcf generated with snpEff
** correction of roxygen issue
** update and editing of the vignette

Changes from Version 2.1.1 to 2.1.2 [March 22 2023]

* Fixing a bug in the pooldata2genobaypass and pooldata2genoselestim export function (the last subset is now printed when using thinning option)
* Function pooldata2genobaypass and genobaypass2pooldata have been accelerated by using fread and fwrite from the package data.table
* Fixing a bug in add.leaf function (I is excluded from possible parameter name for admixture rate)
* Slight modification of computeFST function (allow for NA in the computation of SSI and SSP) 

Changes from Version 2.0.0 to 2.1.1 [May 2 2022]

* Adding a new option in vcf2pooldata.R to allow removing SNPs close to indels (min.dist.from.indels)
* Fixing a bug in the heatmap representation of the pairwise-population FST matrix (the default distfun argument is now set to as.dist instead of dist)
* Adding column names (names of population pairs) to SNP-specific object generated by compute.pairwiseFST  
* Fixing some minor bugs in the functions vcf2pooldata and pooldata.subset to properly manage empty objects
* Minor editing of the vignette and citation file

Changes from Version 1.2.0 to 2.0.0 [May 26 2021]

*Major upgrade of the package including the estimation of f2, f3, f3*, f4 and D statistics and the implementation of a wide range of associated inference methods: (i) three-population test of admixture; (ii) four-population test of treeness; (iii) f4-ratio estimation of admixture rates; and (iv) fitting, visualization and (semi-automatic) construction of admixture graphs.
*Inclusion of new objects, functions and methods and optimization of older ones (not detailed here) 
*Fixing a minor issue in vcf2pooldata function (in version 1.2.0, the reported reference allele read count actually corresponded to the counts for the alternate allele: this had no effect on the estimation of Fst)

Changes from Version 1.1.1 to 1.2.0 [November 3 2020]

* Modifications of vcf2pooldata.R using cpp functions interfaced with Rcpp to improve both speed (about 1 order of magnitude on real life tests) and stability. Parallelisation has been disactivated since it did only provide marginal gain (sequential reading of the file which is not performed in parallel being the main limiting factor).
* Modification of the pooldata.subset.R function to allow subsetting the pooldata object according to SNP indexes (in addition to Pools indexes or not). Additional filtering options included in particular to allow for coverage criteria to be applied on a per-pool basis.
* Fixed a bug that lead to NA in positions not covered by any reads in the subset (i.e., it had no consequence on estimation of statistics) in both the pooldata.subset.R and computePairwiseFSTmatrix.

Changes from Version 1.1.0 to 1.1.1 [October 22 2019]

* Fixed an issue with parsing of vcf files generated by FreeBayes
* Fixed an issue with prefix option and changed writing.dir default value in pooldata2genobaypass an pooldata2selestim functions

Changes from Version 1.0.0 to 1.1.0 [June 23 2019]

* Modified vcf2pooldata function to allows parsing of vcf generated by software other than VarScan (e.g., GATK, BcfTools). Format are automatically detected based on the presence of AD (Bcftools like) or AD and RD fields (VarScan like)
* Modified vcf2pooldata and popsync2pooldata to allow parallelization of some parts of the parsing

Changes from Version 0.0.1 to 1.0.0 [September 14 2018]

* Changed notations in the ComputeFst functions for the MoM method to be consistent with Genetics paper (MoM=>Anova and PiD=>Identity)
* Fixed an issue with missing data in computeFST function (for the MoM method)
* Function pooldata.subset added
* snpdet matrix with dummy values added to the pooldata object created by genobaypass2pooldata and genoselestim2pooldata functions for compatibility with other functions
* Minor corrections