You are here

News / Change Log

kamvarz's picture


poppr 2.0.0


  • poppr has moved to version 2.0 due to adegenet's recent update. The hierarchy slot introduced in version 1.1 is now being moved to adegenet and renamed strata. For maximum backwards compatibility, all of the hierarchy methods still exist, but they are deprecated and will print a warning with the proper function to use. If you were accessing the hierarchy slot without using the *hierarchy() methods, your code will fail as the hierarchy slot now should only contain a formula object.


  • poppr now imports elements from dplyr and shiny
  • As required as of 2015-06-29, poppr now explicitly imports: stats, graphics, grDevices, and utils


  • poppr now suggests the cowplot, poweRlaw, and polysat packages.


  • A data set called Pram containing SSR genotypes from the Sudden Oak Death pathogen Phytophthora ramorum (Kamvar, 2015)


  • refreshing!


  • The default plot for the index of association will now be a single histogram. The user has the option to visualize the standardized index of association (index = "rbarD", default) or the classic index of association (index = "Ia"). If the user uses the function ia with the argument valuereturn = TRUE, then the resulting object can be plotted with the plot function.
  • The function poppr will now plot all populations in a single faceted plot instead of one plot per population.
  • aboot and bruvo.boot will now be able to utilize any function to generate trees (suggested in issue #18).
  • The mlg slot in the genclone object can now optionally hold an MLG class object. This object will contain different definitions of multilocus genotypes, allowing the user to switch between observed, custom, and mlgs defined given a genetic distance threshold.
  • minimum spanning network functions gain the ability to include reticulations using the option include.ties
  • minimum spanning network functions gain the ability to collapse multilocus genotypes by genetic distance with the option threshold.
  • poppr.amova gains the ability to filter multilocus genotypes before calculation.
  • informloci gains the argument "MAF", which allows the specification of a minor allele frequency cutoff in addition to the cutoff argument. Examples have been updated.
  • aboot can now take genlight objects.
  • poppr.msn and plot_poppr_msn can now take genlight objects.
  • plot_poppr_msn now gives users the option to exclude the legends.
  • Default plotting for mlg.table will no longer produce one plot per population. It will now produce a single ggplot object for all populations. Note that the bars are no longer colored by count.
  • poppr will no longer calculate "Hexp". Instead, Simpson's index will be calucluated, but the old index can be retrieved by using (N/(N - 1))*lambda.
  • poppr can now take any statistic that can be caluclated from a table of multilocus genotype counts.
  • genind2genalex gains the ability to selectively write different strata.


  • mlg.filter will contract multilocus genotypes given a genetic distance and threshold using one of three algorithms. It can report statistics such as the multilocus genotypes returned, the number of samples within each multilocus genotype, the thresholds at which multilocus genotypes were collapsed, and the genetic distance matrix that represents the new multilocus genotypes.
  • filter_stats will show you graphical output of all the algorithms in mlg.filter.
  • cutoff_predictor will predict the cutoff threshold from mlg.filter.
  • bitwise.dist can efficiently calculate absolute genetic distance for genlight objects.
  • mll "multilocus lineages" is a new replacement for mlg.vector which gains the functionality of selecting the multilocus genotype definition from the mlg slot.
  • nmll counts the number of multilocus lineages
  • mll.custom allows the user to define custom multilocus genotypes.
  • mll.levels allows the user to edit the names of custom multilocus genotypes.
  • poppr_has_parallel will return TRUE if poppr was built with OpenMP parallel library.
  • win.ia calculates windows of \bar{r}_d along genlight chromosomes.
  • samp.ia calculates \bar{r}_d for genlight object by randomly sampling a user-defined number of SNPs.
  • test_replen will test repeat lengths of microsatellite markers for consistency.
  • fix_replen will fix inconsistent repeat lengths for microsatellite markers.
  • diversity_stats returns a matrix containing diversity statistics. Defaults to 4 found in poppr, but can be extended to any statistic that can be calculated on a vector of MLG counts.
  • diviersity_boot will bootstrap a MLG matrix over the statistics specified for get_stats. Can also perform rarefaction bootstrap.
  • diversity_ci will calculate and plot confidence intervals for bootstrap resampling of an MLG matrix. This includes rarefaction to the smallest sample size.
  • imsn provides an interactive shiny interface for construction of minimum spanning networks.
  • pair.ia will calculate the index of association for pairs of loci and plot heatmaps.


  • snpclone is an extension of the genlight object that acts very much like genclone. It contains an mlg slot.
  • MLG is an internal class that lives inside the mlg slot of snpclone and genclone objects. It allows the user to easily switch between multilocus genotype definitions.


  • I was going too fast to count.

poppr 1.1.5


  • Fixed internal bug for fix_negative_branch when only one branch had a negative edge.
  • Fixed bug in diss.dist where a single locus would return an error.
  • Fixed bug in poppr.amova where a single locus would return an error due to repool_haplotypes.
  • Fixed bug from the future! mlg.table will now return a matrix all the time (Fix #25).

poppr 1.1.4


  • Fixed an internal bug that fails only on Windows OS.

poppr 1.1.3


  • new arguments to plot_poppr_msn to allow for easier manipulation of node sizes and of labeling
  • read.genalex can now take read text connections as input. Addresses issue #8
  • users can now specify cutoff for missing values in aboot


  • Fixed issue where monomorphic loci would cause an error in recode_polyploids
  • Fixed logical error that would cause The infinite alleles model of Bruvo's distance to inflate the distance. (Found by Michael Metzger. Addresses issue #5).
  • AMOVA can now take subset genclone objects (Addresses issue #7).
  • in mlg.table, the mlgsub argument will now subset by name instead of index (fixed in #7).
  • Fixed issue for neighbor-joining trees where the internal function to fix negative branch lengths was accidentally shuffling the corrected branches. Addresses issue #11.
  • diss.dist can now be used with aboot


  • info_table will print a discrete scale as opposed to colorbar when type = "ploidy"
  • attempted to make model choices for Bruvo's distance more clear in the documentation

poppr 1.1.2


  • Fixed memory allocation bug (Further addresses issue #2).
  • Memory allocated in C function bruvo_dist is now properly freed.

poppr 1.1.1


  • Fixed bug where the loss and add options for Bruvo's distance were switched.
  • Fixed illegal memory access error by UBSAN. Made memory management of internal C functions more sane. (Addresses issue #2).
  • Fixed directional quotes and em-dashes produced by Mavericks (Addresses issue #3).

poppr 1.1.0


  • Polyploids with ambiguous genotypes are now supported in poppr. See documentation for recode_polyploids for details.
  • Calculations of Bruvo's distance now features correction for partial missing data utilizing genome addition and genome loss models as presented in Bruvo et al. 2004.
  • diss.dist now has options to return raw distances and a matrix instead of a dist object.
  • read.genalex now has the option to import as a genclone object. This is the default action.
  • poppr.all will be able to analyze lists of genind or genclone objects.
  • ia now has the argument valuereturn which will return the sampled data.
  • [bruvo,poppr].msn functions now give the user the choice to show the graph.
  • bruvo.boot has a cleaner plot style.


  • The genclone object is a new extension of the genind object from adegenet. This object contains slots containing population hierarchies and multilocus genotype definitions and will work with all analyses in adegenet and poppr.


  • [get,set,name,split,add]hierarchy - functions that will manipulate the hierarchy slot in a genclone object utilizing hierarchical formulae as arguments for simplification.
  • setpop will set the population of a genclone object utilizing model formulae regarding the hierarchy slot.
  • as.genclone will automatically convert genind objects to genclone objects.
  • is.genclone checks the validity of genclone objects.
  • poppr.amova will run amova on any hierarchical level. This also includes the feature to run amova on clone censored data sets. It utilizes the ade4 version of amova.
  • info_table will calculate missing data per population per locus or ploidy per individual per locus and gives the user the option to visualize this as a heatmap.
  • locus_table will calculate diversity and evenness statistics over all loci in a genind or genclone object.
  • *.dist functions will calculate Nei's distance, Rogers' Distance, Edwards' Distance, Reynolds' Distance, and Provestis' Distance.
  • aboot will allow the user to create bootstrapped dendrograms for ANY distance that can be calculated on genind or genpop objects.
  • plot_poppr_msn will plot minimum spanning networks produced with poppr.
  • private_alleles will give information about the presence of private alleles within a genind or genclone object.
  • recode_polyploids will take in a polyploid genind/genclone object (with missing alleles coded as extra zero-value allele) and recode them to have frequencies relative to the observed number of alleles.
  • genotype_curve will create a genotype accumulation curve for increasing number of loci.
  • will return a list indicating the samples belonging to a specific multilocus genotype.


  • Pinf - a data set of 86 isolates from different populations of the late blight pathogen, Phytophthora infestans. Provided by Erica Goss
  • monpop - a large data set of 694 Monilinia fructicola isolates from a single orchard over three years. Provided by Sydney E. Everhart


  • Not really.


  • poppr no longer depends on pegas.
  • ade4 and reshape2 are now explicitly required.


  • default shuffling algorithm has been implemented in C to increase speed.
  • output of the mlg functions are now represented as integers to decrease their size in memory.
  • mlg.matrix is now calculated faster utilizing R's internal tabulating capabilities.
  • The function poppr will no longer return rounded results, but rather is printed with three significant digits.


  • Added unit tests.
  • The poppr user manual has been shortened to only include instructions on data manipulation.
  • A new vignette, "Algorithms and Equations" gives algorithmic details for calculations performed in poppr.

poppr 1.0.7


  • Updated README to include link to poppr google group.


  • Made last bug fix more stable (corrected on ape side).

poppr 1.0.6


  • Fixed bug for users who have downloaded ape version 3.1 or higher where bruvo.boot would throw an error.


  • Updated citation information.

poppr 1.0.5


  • The default shuffling algorithm for calculating the index of association has changed from multilocus-style sampling to permutation of alleles. All of the 4 methods are available, but new assignments are as follows: Method 1: permute alleles, Method 2: parametric bootstrap, Method 3: non-parametric bootstrap, Method 4: Multilocus-style sampling. Previously, Multilocus was 1 and the rest followed in the same order. There should be no compatibility issues with this change. Functions affected: ia, poppr shufflepop


  • Bootstrapping algorithm for bruvo.boot function was not shuffling the repeat lengths for each locus resulting in potentially erroneous bootstrap support values. This has been fixed by implementing an internal S4 class that will allow direct bootstrapping of the data and repeat lengths together.
  • An occasional error, "INTEGER() can only be applied to a 'integer', not a 'NULL'" in bruvo.boot or bruvo.dist fixed.


  • Changes to bruvo.boot allow for ever so slightly faster bootstrapping.


  • Permutations for I_A and \bar{r}_d are now visualized as a progress bar as opposed to dots.

poppr 1.0.4


  • A previous error where bootstrap values greater than 100 were reported from bruvo.boot on UPGMA trees has been fixed.
  • Fixed correction of negative branch lengths using Kuhner and Felsenstein (1994) normalization for NJ trees.


  • github repository for poppr has changed from to

poppr 1.0.3


  • Optimized internal sampling function to run up to 2x faster.
  • Utilized rmultinom function to increase speed of bootstrap sampling methods for shufflepop and ia.


  • Function informloci will remove phylogenetically uninformative loci.


  • Now importing specific functions from igraph and ape due to dependency issues.
  • Removed igraph, ape, ggplot2, and phangorn form "Dependencies", but keeping them in "Imports".


  • read.genalex will no longer insert an "X" in front of loci with numeric names.

poppr 1.0.2


  • Fixed bug in diss.dist function that would return an inflated distance for haploids.


  • Added explanation for the index of association in poppr_manual.
  • Expanded installation section to include installation instructions from github.


  • internal permutation algorithm no longer lists permutations in reverse order

poppr 1.0.1


  • Algorithm for the index of association was updated to increase speed.


  • Removed unnecessary rounding factor for missing data in bruvo.dist.
  • Corrected handling of duplicate entries for read.genind.
  • Input values that are not multiples of the specified repeat length for Bruvo's distance are now rounded (as opposed to being forced as integers).


  • Vignette updated for aesthetics and to reflect algorithmic changes.

poppr 1.0.0


  • Poppr has been confirmed to work on Linux, Mac, and Windows systems with R 3.0.0.
  • Vignette poppr_manual now has cross-references to different sections.
  • Vignette poppr_manual is quicker loading.


  • removed alpha channel from plot for resampled values of I_A and \bar{r}_d due to warnings.

poppr 0.4.1


  • getfile has a new argument, "combine", which will automatically add the path to the list of files, so they can be read without switching working directory.
  • information printed to screen from missingno and mlg.crosspop will now be wrapped to 80 characters.


  • poppr will now be able to correctly recognize GenAlEx files with both geographic and regional data.
  • calculation of the index of association on P/A data with missing values will no longer return an error.

poppr 0.4


  • mistake in Bruvo's distance where it did not correctly check for ploidy level was fixed.
  • read.genalex will be able to correctly distinguish between SNP and AFLP data.
  • read.genalex can now correctly recognize regional formatting without an extra column.


  • read.genalex will now be able to take in a file that is formatted with both regional and geographic data.
  • genind2genalex can now export xy coordinates into the GenAlEx format.
  • poppr_manual vignette now contains images of example GenAlEx files.


  • rootrot2.csv is an example of a GenAlEx file formatted with regional data.


  • function for guessing repeat lengths for Bruvo's distance moved into internal file.
  • redundancy in read.genalex was removed.
  • changed instructions in README

poppr 0.3.1


  • read.genalex will now give a warning whenever the input file is not comma delimited.

poppr 0.3


  • poppr.msn will draw a minimum spanning network for any distance matrix derived from your data set.


  • vignette now has sections describing poppr.msn, diss.dist, greycurve, and a section discussing how to export graphics.


  • The graphs output by poppr and ia will now display \bar{r}_d instead of \bar{r}_D.
  • bruvo.boot now has a dedicated quiet argument.

poppr 0.2.2


  • index of association distributions will now feature a rug plot at the bottom as a better way to visualize the distribution of the index of association from the shuffled data sets.

poppr 0.2.1


  • diss.dist will produce a distance matrix based on discreet distances.
  • greycurve will produce a grey scale adjusted to user-supplied parameters. This will be useful for future minimum spanning network functions.


  • bruvo.msn can now adjust the edge grey level to be weighted toward either closely or distantly weighted individuals.
  • bruvo.msn will now return a list giving the user the graph with all of the color, label, and weight properties so that they can plot it themselves. The legend arguments are also returned.


  • fixed shufflepop so that it will now shuffle PA markers with a specific method
  • fixed warning message mistakes in clonecorrect function.

poppr 0.2


  • Added NEWS file and will now be incrementing version number (3/15/2013)

poppr 0.1

  • First development version of poppr (2012 - 3/2013)