• Use unnamed formula argument in aggregate() to account for change in R.
  • DOUBLE_EPS -> DOUBLE_EPSILON.
  • Modified citation to point to Heredity.
  • Adds notes about the necessity of using adaptive priors in ldfast().
  • Includes a win argument in ldfast(), which implements the moment-based LD correction along a sliding window.
  • Fixes a bug where monomorphic SNPs were causing errors when type = "Dprime" was selected in ldfast(). Now we just return NA’s for LD with monomorphic SNPs.
  • Uses the complete reference of Gerard (2021) doi:10.1111/1755-0998.13349.
  • Removes ldfast_old() and ldfast_calc(), which were not used in any exported functions, because these functions had memory issues, detected by valgrind.
  • Added ldfast(), a new LD estimation approach based on sample moments of marginal posterior genotype moments.
  • Unlike ldest(), mldest(), and sldest(), the new approach implemented in ldfast() is scalable to genome-wide applications, as these new estimators can be calculated in linear time in the sample size.
  • Citation of MLE approach points to MER article.
  • I have changed the terminology from “gametic LD” to “haplotypic” LD, and so all instances of “gametic” have changed to “haplotypic”. A breaking change is that all options that were "gam" are now "hap".
  • Fixed an issue where the title in plot.lddf() was being cut off.
  • Added a reference to the preprint where the methodology is developed.
  • Updated the vignette to also take a user through uploading a VCF file into R using the VariantAnnotation package. We also provided examples on formatting genotype likelihoods from updog and fitpoly.
  • Initial release of package.