- quick-N-mix (by Matthew Parker, Lloyd Elliott, Laura Cowen and Jiguo Cao). Fit N-mixture models with large N using asymptotic assumptions.
- agent (by Winfield Chen and Lloyd Elliott). A new compression style for population genetics, with genotype-major univariate multi-phenotype genome-wide association study.
- TreeDiagram (by Wendy Wang and Lloyd Elliott). A CRAN package for visualizing cuts for either axis-align or non axis-align tree methods (e.g. decision tree, random tessellation process).
- peaks (by Winfield Chen). Software to identify clusters of hits among many phenotypes in a genome-wide association study.
- kgen (by Lloyd Elliott). Software to efficiently compute inverses of genetic similarity matrices for partially observed multi-phenotype data.
- tess19 (by Shufei Ge and Lloyd Elliott). An implementation of the random tessellation forest.
- HDPStructure (by Maria De Iorio, Lloyd Elliott, Stefano Favaro and Yee Whye Teh). A Bayesian nonparametric generalisation of the Structure2 model for genotype imputation and admixture inference.
- bgenie (by Jonathan Marchini and Lloyd Elliott). Genome-wide association studies on multi-phenotype data. Errata July 2020: 1) There may be a bug involving the joint use of the flags --exc_missing_inds and --covar. Details of this possible bug are provided on the oxstatgen mailing list here. Until this is investigated, I’d recommend performing all covariate deconfounding on phenotype files before passing them to bgenie and not using the flags --exc_missing_inds and --covar, for example by preprocessing based on the code here (note that --covar used without --exc_missing_inds may be okay). 2) For the X chromosome non-pseudoautosomal region, if the codings for genetic male genotypes are rescaled to the range [0, 2], then genetic males will incorrectly contribute twice their ploidy to the MAF and INFO score columns of the output files.
- Mondrian shapes (by Lloyd Elliott). A procedurally generated interactive art project.