v1.5

dispRity v1.5.0 (2020-09-25) between groups

NEW FEATURES

• New function: multi.ace for performing fast ancestral character estimations on multiple matrices (based on castor::asr_mk_model).
• New function: reduce.space, a function to modify trait spaces imported from the moms shiny app. This function comes with a new reduction algorithm: the "evenness" algorithm for flattening the curve (thanks to Gavin Thomas for the suggestion).
• New function: test.metric (and associated plot, print and summary functions), to apply the reduce.space function on a specific space and metric to test whether a metric is picking up specific changes in trait space.
• the dispRity function can now use "between.groups" metrics to calculate disparity between groups rather than within groups. The make.metric function is now modified to allow detection of metrics that can be applied between groups.
• New metric: group.dist, a dimension level 1 metric for between groups that measures the distance between two groups. By default, this is the minimum distance but the function takes the probs argument allowing the distance to be between, says, the 95% CI (probs = c(0.025, 0.975))) or between the centroids (probs = c(0.5)).
• New metric: point.dist, a dimension level 2 metric for between groups that measures the distance between the rows in matrix to a point in matrix2. That point is the centroid by default but the point argument can take any function.

MINOR IMPROVEMENTS

• The dispRity package now depends on R (>= 4.0.0).
• Many updates to the dispRity manual.
• Many minor speed improvements across the package
• Simplified syntax for the internal plot.dispRity S3 methods (for a potential ggpRity?). These changes should not be apparent at the user level but see the two removed options below:
• removed option in plot.dispRity: the chrono.subsets option (TRUE/FALSE) has now been removed. The time-slicing-ness is now automatically detected or can be specified by the user normally through xlab.
• removed option in plot.dispRity: the ylim, ylab, xlab and col options have now been removed. They are now handled through ... as normal generic plot(...) arguments.
• removed option in plot.dispRity: the elements.pch option has now been removed. The pch of the plotted elements can now be passed like other options directly to elements (e.g. elements = list(pch = 15)).
• removed option in plot.dispRity: the dimensions, matrix, nclass and coeff options have now been removed. Any options for dual class plots (randtest, dtt, model.test, type = "preview" etc...) are now handled through the generic specific.args argument.
• New option in dtt.dispRity: scale.time allowing to scale the time axis (like in geiger::dtt) or not.
• when plotting chrono.subsets dispRity objects, the x label ticks are now rounded if possible (for nicer looking plots!).
• when using automatic chrono.subsets time slices, the name of the time slices (their age) is now rounded for aesthetics.

BUG FIXES

• Updated all calls to the Claddis package to match version 0.6 (thanks to Graeme Lloyd for fixing this one!)
• Updated Claddis.ordination function to be compatible with the new ape version 5.4 (thanks to Emmanuel Paradis for the pointing that out).
• Fixed a bug in chrono.subsets where ladder trees with no tip branch lengths an method = "continuous" option would get stuck in an infinite loop.
• Fixed a bug in chrono.subsets where the "*.split" methods would bug if the last slice is through a single edge.