- First release.
- Simulation of additive genetic effects now uses
MCMCglmm::rbvinternally. It is substantially faster! - Removed nadiv dependency until it back on CRAN, so temporarily doesn't simulate dominance or epigenetic effects internally. This can still be done by including matrices generated from nadiv as covariance structures
- Bug fixes with simulating additive genetic effects, and added index_link argument to help with indexing in the model argument. This allows new factors in the data structure to be made which are indexed by other factors. Previously simulating maternal genetic effects didn't work properly, because of the indexing, and this solves that problem.
- New survival sampling functionality. If you simulate binomial data with a age structure, then you can sample so that you only have the ages up to (and including) the first 0 or 1 (depending on whether you are simulating survival or mortality)
- Bug with 'fixed' factor is fixed. Now if the levels in the data structure have names, the same names have to be specified in the parameter list. The order doesn't matter (this will be matched internally), but the names have to match exactly. If the factor levels in the data structure are numeric then these do not need to match; the order of the effects will match the order of the numeric factor levels.
- Added 'int_names' argument to
make_structurefunction, that allows combination/interaction terms to be named.
- Added cloglog link function to
simulate_population - Added functions
exp2latandlat2exp, which convert means and (co)variances between normal and log normal distributions, and so can be used to transform between latent and expected scales with GLMs - Fixed phylogenetic effect simulation
- Optimising sparse matrices to speed up simulating with
cov_str - Updating internal error checking