shawnlaffan/biodiverse: Version 4.0

Executables/Binaries Executables/Binaries are available below for Windows, Macintosh and Linux (Ubuntu). There are two Mac variants. If you are using a Mac with an M1 chip then download the mac_arm zip file. If you are using a Mac with an Intel chip then download the mac_intel zip file. Installation instructions are available at https://github.com/shawnlaffan/biodiverse/wiki/Installation Changes For the full list of issues and changes leading to the 4.0 release, see https://github.com/shawnlaffan/biodiverse/milestone/17 A series of blog posts goes into more detail about several of the changes: http://biodiverse-analysis-software.blogspot.com/search/label/Version4 General As of 3.99_005 the Windows executable files are digitally signed. This will avoid OS warnings when files are downloaded. Analyses CANAPE Biodiverse now automatically generates and plots CANAPE results for any analysis that contains the phylogenetic endemism and relative phylogenetic endemism indices. Issue 819. More details are in this blog post. The PE and RPE calculations have been sped up. PR 831. Cluster analyses Cluster analyses will now lump all zero distance matrix pairs in an initial pass where possible. This avoids needless tie-breaker calculations when data sets have large sets of groups with the same label assemblages. Issue 760 Cluster and Region Grower analyses are no longer rebuilt by default under the randomisations, speeding up any randomisations that include them. Comparison of per-node calculations is still done, though, as this does not use the rebuilt tree. There is a user visible option in the GUI to enable the previous behaviour if it is needed. Issue 765 Cluster and region grower analyses now warn and provide an informative error message when there is only a single group in the basedata. Issue 782 Randomisations Randomisations now also generate z-scores of the observed values against the random distribution. Issue 764. More details are in the blog post. The Independent Swaps algorithm is now implemented. Issue 772. More details are in the blog post. Calculations A local PD calculation, "Phylogenetic Diversity (local)", now allows the calculation of PD and PD_P to the last common ancestor, as opposed to the root node. Issue 762 A new calculation, "Last shared ancestor properties", provides details about the last shared ancestor for a set of labels, for example the length from its tipwards end to the root node. Issue 763 MPD, MNTD, NRI & NTI: The MPD and MNTD calculations are now substantially faster. The NRI and, for ultrametric trees, NTI calculations are orders of magnitude faster as they use expected values calculated using the exact methods described in Tsirogiannis et al. (2012) and Tsiriogiannis et al. (2014). There is also no longer a performance penalty under the randomisations for outputs containing NRI and NTI indices when richness scores are held constant (which is the default). This is because the expected values for a given tree are re-used across analyses. Issue 786, Issue 789, Issue 790. More details are in the blog post. The taxonomic distinctness calculations have been moved to the BiodiverseX namespace and so are not available in the GUI by default. Issue 751. The calculation of median and percentile values for the element properties have changed. Issue #798. More details in the blog post. The phylogenetic distincness variance (VPD) is now calculated with the mean and other related statistics. The net VPD and its component indoces can also be calculated to assess significance of the net VPD against random resampling. Issue #811 See blog post Spatial Conditions New condition sp_richness_greater_than() allows the neighbour set to be defined as the set of groups with a richness exceeding a threshold. Issue 783. New condition sp_redundancy_greater_than() allows the neighbour set to be defined as the set of groups with a sample redundancy exceeding a threshold. New conditions sp_point_in_cluster() and sp_points_in_same_cluster() that can be used to model polygons from clusters in a Cluster or RegionGrower analysis without first needing to export them to a shapefile and any subsequent processing to extract the relevant parts. Issue 803. More details in the blog post. Imports Group properties can now be directly imported from rasters. This makes it much easier to analyse environmental patterns in relation to taxonomic and/or phylogenetic ones. Issue 761. More details are in the blog post Exports Cluster and RegionGrower analyses can be exported to shapefile format in a grouped form. This models the cluster display where multiple sub-clusters are coloured. The output file is somewhat awkward and needs further processing, which is why the sp_points_in_same_cluster() spatial condition was developed (see above). Issue 757. More details are in the blog post Trees A new option has been added to merge only-child ("knuckle") nodes with their parents. This is probably most useful after trimming a tree but can be applied to any tree. Issue 814. More details are in the blog pos