dispRity v1.6.8 (2022-01-24)

NEW FEATURES

New data function: select.axes for selecting and analysing the number of axes required to contain an arbitrary amount of variance.
New utility function: randtest.dist for measuring the distance between the observed statistic and a specific quantile of the simulated statistic (thanks to Frane Babarovic for the inspiration).

MINOR IMPROVEMENTS

Formalised the grouping logic for custom.subsets and select.axes. This can create some minor user level changes namely: warning messages for empty subsets now correctly mentions “subsets” (rather than subsamples); groups with incorrect elements are now always flagged as errors (rather than just ignored). The changes at the developer level is that the logic is now made smoother and exported in custom.subsets_fun.R.
Added a function.index.csv list (and updater) to help developers find internal functions locations easily.
Restricted the type-I error inflation warning message in test.dispRity to only occur when using a test of class "htest".
Continuous Integration has been moved from Travis-CI to GitHub Actions.

BUG FIXES

Removed warning in dispRity when selecting a specific number of dimensions (old warning artefact).
Fixed bug in plot.dispRity when using type = "preview" on bootstrapped data and for type = "box" when the subsets to plot are from different sizes (now plots all the data correctly).
Fixed bug when using chrono.subsets with "continuous" method a FADLAD data containing only node values (now correctly taken into account; thanks to Peng-Wei Li for noticing it) and when using chrono.subsets with "gradual.*" models on empty subsets.
standardGeneric functions are now correctly interpreted as functions throughout the package.
Fixed bug when plotting level 1 disparity metric results without bootstrapped (observed = TRUE is now used as the default).
Fixed bug when plotting test.metric plots with save.steps options with more than two types of shifts.
Fixed bug with null.test which is now correctly managing the number of dimensions inherited from dispRity objects (thanks to Alex Slavenko for spotting this one and the two above).
Fixed bug when using level 2 dimension metrics on unidimensional data (the metric is now detected as a level 2 correctly; thanks to Catherine Klein and Rachel Warnock for noticing that one).
Update internal use of is(data, c("array", "matrix")) to is.array(data) for R 4.1.2.

dispRity v1.6.0 (2021-04-16) dispRitree

NEW FEATURES

New metric: projections that allows to measure elements’ projection on an arbitrary axis (or their distance from this axis with measure = "distance").
New metric: projections.tree that allows to measure elements’ projection on axis between elements of a given tree.
New metric: edge.length.tree the edge length from each element given a tree (with the option to.root = TRUE/FALSE to measure the edge length from the element to the root of the tree (default = TRUE) or the nearest ancestor (FALSE).
You can now save the shifts results in test.metric with save.steps and then visualise them with plot.dispRity along side the disparity metric test results.
New utility function n.subsets to directly get the number of subsets in a dispRity object.
New statistical test: randtest.dispRity that is a wrapper for ade4::randtest applied to dispRity objects (not dissimilar from null.test).
Six more demo datasets have been added to the package! These datasets are the ones used in Guillerme et al. 2020 and published originally in Beck & Lee 2014 (that one was originally the only demo dataset in the package), Wright 2017, Marcy et al. 2016, Hopkins & Pearson 2016), Jones et al. 2015, Healy et al. 2019. Thanks to all these authors for their open science work!
dispRity objects now have a reserved $tree component that contain any number of trees attached to the data. This allows any function to use the reserved argument name tree to extract directly the relevant tree from the dispRity object, for functions like chrono.subsets or metrics like ancestral.dist! To help manipulate the tree component of the dispRity object, you can now use the new utility functions add.tree, get.tree and remove.tree.

MINOR IMPROVEMENT

Reverted R version requirement from 4.0 back to 3.6 following Joseph Brown’s issue and fix.
reduce.space "random" algorithm now outputs a named logical vector (like the other algorithms!).
remove the "only" text when printing dispRity objects that contains “only” matrices (even though that can be 10k matrices!).
added a dedicated behaviour to summary.dispRity for "dispRity" "randtest" objects to output “ready-to-publish” result tables.
some error messages have been updated to be slightly more useful.
added the estimation.details argument to multi.ace allowing to also return specific arguments from the ancestral states estimation (thanks to Armin Elsler for the suggestion).
Added new option inc.nodes to clean.data whether to check if the nodes in the tree match the labels in the matrix.
make.metric with the option silent = TRUE now outputs a list of info rather than only the level of the metric. You can reproduce the old behaviour using make.metric(..., silent = TRUE)$type).
Fixed bug in plot using preview when the given argument pch did not match the number of groups (the different pch arguments are now used correctly).
Completely revamped the ancestral.dist metric. The function is now much faster and much easier to use (due to the new dispRity object structure). The options nodes.coords has been removed and the option full is now changed by to.root. If you still really want to use the older version of ancestral.dist using ancestral.dist.deprecated though.
The dimensions option throughout the package (e.g. in the dispRity function) can now also be a vector of dimensions to take into consideration (e.g. c(1,2,5)).

BUG FIXES

chrono.subsets now automatically detects the number of digits to round for the internal time slicing functions (thanks to Mario Corio for finding this one).
Fixed bug in test.metric plots that now display correctly the “top” and “bottom” changes for the “position” shift.
Fixed bug in test.metric plots that now display the R^2 values correctly.
Fixed bug in tree.age when the tree tips/node labels vector is longer than the actual number of tips/nodes in the tree.

DEPRECATED AND DEFUNCT

Removed former version of ancestral.dist (see NEW FEATURES above).
Removed node.coordinates function (no replacement; you must use a package version prior 1.5.10 to use this function).
Removed get.ancestors function (no replacement; you must use a package version prior 1.5.10 to use this function).

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.
Fixed a bug in dispRity where some 1D matrices could loose their class for certain metrics.

dispRity v1.4.0 (2020-05-05) getting faster

NEW FEATURES

New metric: angles, a dimension level 2 metric that measures the angle of the main axis of each dimension in a matrix (in slopes, angles or degrees).
New metric: deviations, a dimension level 2 metric that measures the deviation of each element from a hyperplane.
Completely rewritten char.diff function. It now uses a way faster bitwise comparison architecture and comes with different distance methods as well as modular optional arguments on how to treat various special tokens ("?", "-", "&", etc.). This also allows many more distance methods (now including "hamming", "manhattan", "comparable", "euclidean", "maximum" and "mord").
all dispRity functions can now intake a single "matrix" or a "list" of matrices with the same row names and dimensions. The disparity is then calculated directly on all the matrices and summarised as before through summary.dispRity. This option can be used to add uncertainty to disparity calculations. For example in chrono.subsets you can now provide a list of trees and a list of associated ancestral state estimates; or for custom.subsets you can provide a list of matrices with different values representing different estimations of the traits.

BUG FIXES

update reduce.matrix to work with vegan::vegdist version 2.5-6 (thanks to Jari Oksanen for the fix).
updated class evaluations throughout the package for R version 4.0.0: class(.) == * is now is(., *).
updated ... argument bug PR#16223.
In make.metric the argument ... is now ignored if any names(...) is "tree" or "phy".
fixed bug in neighbours and span.tree.length when feeding “distance” like metrics (thanks to Ashley Reaney for finding that one).

MINOR IMPROVEMENTS

greatly improved speed of chrono.subsets with method = "continuous (now > 1000 times faster!).
minor warning message fix for plot.dispRity with time slices.
removed paleotree package dependency (though the links to this excellent package are still there!).
increased R version requirement to 3.5.
... in summary.dispRity are now directly passed to cent.tend (e.g. na.rm = TRUE).
added some time improvements in several phylo functions based on the castor package.
updated all the package demo data to fit the new dispRity object architecture (see above). Note that this might effect the exact results of calculations using these demo datasets.
you can now specify the dimensions of the matrix to make a disparity metric in make.metric through the data.dim option.
metrics passed in dispRity are now tested using the input data dimensions.
chrono.subsets with multiple trees now stretches the root edges length to match the oldest tree.

dispRity v1.3 (2019-08-15) many metrics

NEW FEATURES

New metric: displacements, a dimension level 2 metric that measures the position of elements in space (i.e. their distance from the centre relative to their distance to their centroid).
New metric: neighbours, a dimension level 2 metric that measures the distance from an element to its neighbour (e.g. the nearest neighbour, the furthest, the median, etc.).
New metric: quantiles, a dimension level 2 metric that measures the nth quantile range per axis (a good alternative to the ranges function!).
New metric: func.eve, a dimension level 1 metric that measures the functional evenness (i.e. the spread along the minimum spanning tree; from Villéger et al. 2008).
New metric: func.div, a dimension level 1 metric that measures the functional divergence (i.e. the ratio of deviation from the centroid; from Villéger et al. 2008).
Updated metric: span.tree.length now outputs the length of each edges (c.f. the sum of the length) and becomes a level 2 metric.
The chrono.subsets can now take multiPhylo objects for slicing through multiple trees at once!
New utility function: reduce.matrix for optimising data overlap in a matrix with missing data.
New utility function: slide.nodes for sliding specific nodes on a tree.
New utility function: remove.zero.brlen for stochastically removing zero branch lengths on a tree (using the slide.nodes function).
New argument in plot.dispRity: the type argument can now be "preview" to have a glimpse at two of the dimensions of the trait-space.
The Claddis.ordination can now directly take a matrix’s path as input (leaving the function to read and transform the matrix into Claddis format. The function can thus now also be used to convert matrices into Claddis format.
Added a “Other functionalities” section to the manual describing miscellaneous functions.

MINOR IMPROVEMENTS

centroids and ancestral.dist functions can now take the method option for "euclidean" or "manhattan" distances.
All functions methods selection have now been sped up using switch.
Error messages in dispRity are more verbose when input the wrong metric(s).
scree option in space.maker does not require to sum up to one anymore.
cor.matrix option in space.maker does not require to have a valid Choleski decomposition (an approximation is used instead).
Updated all tests and functions to be compatible with R 3.6.
Fixed bug in clean.data that did not output dropped tips correctly when applied on multiPhylo objects.
Improved error messages in chrono.subsets for funky time slices/bins (e.g. with negative values).
Speed improvements for the time.slice function.
Better internal handling of distance matrices for the disparity metrics.
Most functions handles NA as na.rm or na.omit.

dispRity v1.2.3 (2019-03-12)

NEW FEATURES

dispRity objects now contain a metric argument (if a metric was applied). This argument can now be recycled by the appropriate functions (e.g. in null.test).

MINOR IMPROVEMENTS

plot.dispRity argument observed can now take a list of arguments to be passed points().
boot.matrix now makes a error warning message when bootstrapping distance matrices (as suggested by Dave Bapst!).
geomorph.ordination can now be used to simply create coordinates matrices (no ordination!) with ordinate = FALSE argument.
better internal handling of error messages.
faster eigen value estimations in ellipse.volume when the argument is left missing.
removed internal handling of the Claddis.ordination function. This function now uses the brand new version of the Claddis package on CRAN (0.3).

BUG FIXES

plot.dispRity with option "box" now correctly display plot ranges when disparity is an observed distribution.
test.dispRity handles errors messages more efficiently when disparity is an observed distribution.
summary.dispRity handles non-bootstrapped distributions display properly.
geomorph.ordination now converts "character" vectors into "factors".
adonis.dispRity now properly handles complex formulas (with arithmetic signs).
... are now properly handled by internal metric testing functions for more accurate error messages.
char.diff names are now properly protected in the C implementation to comply with new rcheck requirements.

dispRity v1.2 (2018-09-19) model tests

NEW FEATURES

New functions: model.test, model.test.sim and model.test.wrapper for fitting models of disparity evolution through time (with associated manuals, vignettes and S3 methods! Thanks to Mark Puttick).
New argument in boot.matrix: prob for passing probabilities of sampling for specific elements.
S3 print method for objects of class "dtt" and "dispRity" (from dtt.dispRity).

MINOR IMPROVEMENTS

tydiversed most of the error messages.
dtt.dispRity now allows to specify the alternative hypothesis (if nsim > 0).
ellipse.volume can now take an explicit eigen value vector (the eigen values are still automatically estimated correctly for PCO and MDS).
Improved metric checking messages from make.metric when dealing with optional arguments.
Removed cascade of warnings triggered by plot.dispRity.dtt.

BUG FIXES

Corrected char.diff to properly reflect the probability of different splits between characters (thanks to Abigail Pastore).

dispRity v1.1 (2018-03-20) got CRAN

NEW FEATURES

CRAN release 1 with the additional Claddis.ordination function.

MINOR IMPROVEMENTS

Added default Cailliez correction to Claddis.ordination function (with add = TRUE).
Improved test coverage.

dispRity v1.0.3 (2018-03-20) got CRAN

NEW FEATURES

First CRAN release

MINOR IMPROVEMENTS

Registered C symbols properly.
Overall improvement for the S3 methods.
Changed name: scale.dispRity is now rescale.dispRity.
Changed name: merge.subsets is now combine.subsets.
Changed name: time.subsets is now chrono.subsets - time.subsets can still be called as an alias for the same function.

BUG FIXES

Minor bug fixes and typos in various error messages and in the manual.

DEPRECATED AND DEFUNCT

Removed Claddis.ordination function to comply with the CRAN requirement (this function is still live in the GitHub version 1.1).

dispRity v0.5 (2017-12-20) covered with tests

NEW FEATURES

custom.subset can now automatically create clade-based groups if a phylo object is passed to group.
New utility function: extinction.subsets, to get the list to be passed to test.dispRity for testing the effect of extinction.
New test function: dtt.dispRity, a wrapper for geiger::dtt. This version is slower that geiger::dtt but allows any univariate disparity metric!
New test function: adonis.dispRity, a wrapper for vegan::adonis.
New utility function: crown.stem for separating a tree into crown and stem groups.
New disparity metric: span.tree.length the length of the minimum spanning tree.
New disparity metric: pairwise.dist: the element’s pairwise distances.
New disparity metric: radius: the radius of each dimensions.
New disparity metric: n.ball.volume: the n-dimensional sphere or ellipsoid volume.
New argument in time.subsets, model = "equal.split" and model = "gradual.split" that retain the probability of being either the descendant or the ancestor. This probability is passed to boot.matrix .

MINOR IMPROVEMENTS

Changed calls to stats::dist to vegan::vegdist to allow more distances to be passed through methods arguments.
slice.tree can now slice through a single edge.
Various minor speed improvements.

BUG FIXES

Correct behaviour in tree.age to estimate ages for trees with fossils only.

DEPRECATED AND DEFUNCT

Change name throughout the package, subsample is now replaced by subset (e.g. time.subsamples is now renamed time.subsets, data$subsamples is now data$subsets, etc…).
Changed argument in time.subsets, model = "gradual" is now replaced by model = "proximity" and model = "punctuated" is now replaced by model = "random".

dispRity v0.4.1 (2017-11-13)

NEW FEATURES

New disparity metric: ancestral.distance to get the distance from taxa/nodes to their ancestors.
New function: random.circle for generating random circle coordinates (see example in space.maker for creating doughnut spaces!).
New function: get.bin.ages for getting the geological timescale of a tree (based on geoscale).

MINOR IMPROVEMENTS

Added a t0 argument to time.subsamples allowing to set the start age of the first subsample.
Allowing subsamples to contain less than three elements (up to 0!).

BUG FIXES

Fixed fuzzy match issues in slice.tree.

dispRity v0.4 (2017-08-21) user friendly

NEW FEATURES

Entirely rewritten manual (in GitBook)!
New function: Claddis.ordination and geomorph.ordination for automatically ordinating data from Claddis and geomorph packages!
New function: char.diff for calculating character differences and associated plot function (plot.char.diff)
New utility function: merge.subsamples for… merging subsamples.
New utility function: size.subsamples for getting the size of subsamples in a disparity object.
New wrapping functions: dispRity.through.time and dispRity.per.group now runs easy default disparity analysis.

MINOR IMPROVEMENTS

Input ordinated matrices do not need to be of maximum size n*(n-1). Bigger matrices now only trigger a warning.
Added dimensions optional argument to dispRity to overwrite the number of dimensions generated by boot.matrix.
variances, ranges and centroids are now simplified for speed. The optional arguments for data cleaning are now passed to make.metric.
space.maker now allows to approximate the dimensions variance distribution with the scree option.

DEPRECATED AND DEFUNCT

Removed hyper.volume metric for dependencies reasons,
Removed parallel option from boot.matrix (the new architecture is already super fast: <2sec for 5k taxa and 10k bootstraps!).
Changed name: series as a part of dispRity objects is now changed to subsamples throughout the whole package.
Changed name: time.series is now renamed time.subsamples, if dates are provided and method is discrete, this function doesn’t need a phylogeny any more.
Changed name: get.subsamples.dispRity is now renamed get.subsamples.
Modified function: cust.series is now renamed custom.subsamples (to avoid confusion with custard.subsamples!). Its factor argument as been changed to groups and can now take a simple list.

dispRity v0.3 (2017-01-25) dispRity lite

NEW FEATURES

Complete change of the dispRity object architecture (see more here).

MINOR IMPROVEMENTS

sim.morpho can now use model = "mixed" for using both HKY-binary and Mk in characters simulation.
dispRity object utilities are now all grouped under the ?dispRity.utilities manual with appropriate S3 methods.

DEPRECATED AND DEFUNCT

Removed rm.last.axis argument in boot.matrix. It is now replaced by dimensions.
Changed argument in plot.dispRity, type = "lines" is now replaced by type = "line".

dispRity v0.2.1 (2016-11-07) making stuff up

NEW FEATURES

New utility function: merge.time.series for cleaning or merging time series,.
New vignette: dispRity-simulate_data on how to simulate morphological characters in dispRity.
New function: sim.morpho generates morphological matrices.
New function: check.morpho for checking how “realistic” the simulate morphological matrices are.
New utility functions: get.contrast.matrix and apply.inapplicable functions for morphological matrices.

MINOR IMPROVEMENTS

Minor updates for the graphical functions.

dispRity v0.2 (2016-06-15)

NEW FEATURES

New utility functions: pair.plot, scale.dispRity and sort.dispRity.
New function: space.maker for creating some multidimensional spaces!
New disparity metrics: convhull.surface, convhull.volume and hyper.volume.
New disparity test null.test.
New plot.dispRity arguments: density for controlling the polygons density and add for adding plots.

MINOR IMPROVEMENTS

Many updates to the functions manual and vignettes.
Some algorithm are now greatly improved for speed.
Disparity can now be calculated as a distribution (i.e. dispRity, test.dispRity, plot.dispRity, and summary.dispRity can now intake one or more distribution rather than just one or more single values of disparity; whether the data is bootstrapped or not).
dispRity can now intake dispRity objects with level 2 disparity metrics.
boot.matrix and dispRity can now run in parallel.
centroids disparity metric can now use a centroid argument for fixing the centroid point value.
variances and ranges disparity metrics can now intake a k.root argument for scaling the results.

BUG FIXES

Minor functions corrections for specific optional arguments combinations.

DEPRECATED AND DEFUNCT

Removed type_discrete argument in plot.dispRity and type argument can now be:
- continuous disparity curves.
- box for real boxplots.
- lines for the distribution vertical lines.
- polygon for the distribution boxes.

dispRity v0.1.2 (2015-12-01)

NEW FEATURES

New function: get.dispRity for subsampling dispRity objects.
New function: extract.dispRity for extracting disparity results.
New function: test.dispRity for applying tests to dispRity objects.
New function: make.metric for helping creating your very own disparity metric.
New metric: hyper.volume for measuring the morphospace hyper-ellipsoid volume.

MINOR IMPROVEMENTS

metric argument from dispRity can now intake up two three functions (see dispRity.metric and make.metric).
Many improved functions manuals and examples!
Improved vignettes:
- dispRity palaeo demo a quick demo aimed more for palaeobiologist.
- dispRity ecology demo a quick demo aimed more for ecologists.
- dispRity manual for people that want to know the package in details.
- dispRity metrics for explaining how the disparity metric implementation works.

dispRity v0.1.1 (2015-10-08)

MINOR IMPROVEMENTS

plot.dispRity options improved (rarefaction + default).
cust.series can now intake multiple factors columns.
Added example for ecological data.
boot.matrix, dispRity, summary and plot now also include observed values.
plot now has an observed option to plot the observed disparity.

BUG FIXES

Fixed many error/warning messages.

DEPRECATED AND DEFUNCT

Changed taxa to elements.
plot option diversity has been renamed elements.

dispRity v0.1 (2015-10-01)

First release!