Version History

From BEAST Software

Jump to: navigation, search

BEAST Documentation->Version History

  • This version addresses the following issues (" * " indicates critical issue):

Contents

Version 1.6.2 released 23rd August 2011

New Features
        adds a "Beta Diversity" statistic to TreeStat program. This statistic
        takes a tree of two sets of species and computes the percentage of
        the tree length shared by the two taxa sets.

        allows BEAUti 'Guess Dates' dialog to parse out calendar dates from
        the taxon names.

        Updates all native nucleotide likelihood core files into 64 bit.
        For 32 bit machines, the user needs to rename the 32 bit files
        (e.g. for Windows, rename NucleotideLikelihoodCore32.dll into
        NucleotideLikelihoodCore.dll) to replace the current 64 bit files.
Bug Fixes
        Issue 423:	delta exchange operator description is wrong
        Issue 424:	TraitGuesser not extracting match within brackets (braces)
        Issue 434:	BEAUti: Initial value missing from prior dialog
      * Issue 435:	UniformIntegerOperator did not include upper in the value
        Issue 454:	BEAUti: all the taxon sets are deleted when link/unlink tree
        Issue 460:	LogCombiner uses localized decimal point when writing trees
      * Issue 464:	Localized decimal indicator still being used in log files
        Issue 472:	Don't report operator suggestions when autooptimize is on
      * Issue 473:	BEAGLE default options should be CPU/Double
        Issue 474:	Partition table shows 0 taxa for trait partitions
        Issue 478:	TipDates table cell editor large the text font size when
	                 changing the value
      * Issue 479:	Multi-block charset parsing and handling in BEAUTi
        Issue 484:	beagle_scaling command line flag turns on BEAGLE in the
                         absence of any other flags
        Issue 486:	BEAST cannot find NucleotideLikelihoodCore.dll in Windows
        Issue 487:	BEAUti: Wrong error message with poorly constructed priors
        Issue 488:	BEAUti: Poor parsing of Poisson rate and offset in Priors
        Issue 489:	Default Laplace prior mean is 0.01
      * Issue 491:	BEAUti parameter should keep uniform upper and lower
        Issue 492:	Should BEAUti use 'NNN' rather than 'N' for empty alignment,
                         to allow for codon-partitioned models
        Issue 494:	BEAUti *BEAST does not need link tree prior option
        Issue 495:	Tree annotator needs to validate the duplicate taxon name

Version 1.6.1 released 15th October 2010

Bug Fixes
        Issue 398: BEAUti text message overlaps with button
        Issue 399: Binary for TreeStat packaged with BEAST v1.6 has wrong version
                   number and icon
        Issue 403: Starting tree generated with constraints by BEAUti is not correct
                   XML for BEAST
        Issue 404: Make "Median heights" as the default of nodeHeightsCombo
        Issue 405: RandomWalkOperator with reflecting boundary can grind to a halt
      * Issue 407: *BEAST needs to log species tree in { }, otherwise TreeAnnotator
                   keeps coming up with null exception on species tree
        Issue 408: TreeAnnotator crashes with Null Pointer Exception on malformed
                   metacomment
      * Issue 410: "Improper priors" on posterior statistics
        Issue 411: Incorrect validation and error message for random local clock
                   in BEAUti
      * Issue 412: BEAUti generates randomWalkOperator instead of randomWalkIntegerOperator
                   for branchRates.categories
        Issue 413: BEAUti: duplicate class PartitionModelOptions causes model link conflict
        Issue 414: BEAUti: incorrect xml for multi-alignment when given complex
                   combination of clock model and tree model
        Issue 416: Laplace prior parser should use scale same as BEAUti
        Issue 417: BEAUti growthRate related XML causes -INF likelihood
        Issue 418: Set maximun fraction digits for numbers displayed in BEAUti prior panel
        Issue 419: BEAUti RealNumberField should not throw Exception all the time
        Issue 422: BEAUti: add validation to check whether modified prior initial value
                   is between the boundary

Version 1.6.0 released 17th September 2010

New Features
	Random local clock model, see:
	Drummond AJ and Suchard MA (2010) Bayesian random local clocks or one
	rate to rule them all. BMC Biology.

	Discrete trait models, see:
        Lemey P, Rambaut A, Drummond AJ & Suchard MA (2009) Bayesian
        Phylogeography Finds Its Roots.PLoS Comput Biol 5, e1000520.
	http://beast.bio.ed.ac.uk/Tutorials#Phylogeography_tutorials

	Continuous trait models, see:
        Lemey P, Rambaut A, Welch JJ & Suchard MA Phylogeography takes a
        relaxed random walk in continuous space and time. (2010)
        Mol Biol Evol 27, 1877-1886.
        These models are not currently specifiable from BEAUti, but see here:
        http://beast.bio.ed.ac.uk/Tutorials#Phylogeography_tutorials

        Improved user interface for Discrete Trait models in BEAUti.

	Added an 'Include Stem?' option in the Taxon Set table  in BEAUti.
	    This allows the calibration to be specified for the stem of a clade.

	Added the application TreeStat to the BEAST package.

	Can 'drag and drop' files to the data table in BEAUti to import them.

	New 'tool tip' help messages for most user-interface elements in BEAUti

	The name of the XML file is now shown in the title bar of the BEAST window.

	By default, BEAST will no longer overwrite existing log/trees files. This
	    behaviour can be overridden by using the '-overwrite' command line option
	    or the check box in the options dialog box.

	There is a BEAST new command line option '-prefix XXX' which will add the
	    specified text to start of the file names of all log files and trees files
	    for that run. This is useful for batch running of BEAST.

	More warnings are given now about potentially problematic prior choices,
	    in particular the use of improper priors. These are shown in yellow.
	    For some models, the user now has to make an explicit prior choice.
Bug Fixes
  • This version addresses the following issues: 
	Issue 40:  TreeAnnotator produced misleading medians and HPDs when calculated
	           from only 1 tree.
	Issue 170: coalescentLikelihood would sometimes result in a zero initial likelihodo
               when using an UPGMA starting tree
	Issue 235: BEAUti: Unlinked DiscretizedBranchRates for multiple loci have the
	           wrong number of dimensions
    Issue 291: BEAUti: fixed a problem generating a user specified starting tree 
               in XML.
	Issue 298: BEAUti: checks that a user specified tree is rooted and fully
	           bifurcating.
	Issue 300: BEAUti: Simplified the XML for discrete models by using a single
	           'generalSubstitutionModel' parser for all variants.
	Issue 302: BEAUti: Now possible to importing multiple files at once when they
	           have different taxon sets
	Issue 307: BEAUti: missing .txt in tree log using EBSP
	Issue 311: BEAUti: Prior choices for codon partition parameters are now
	           properly generated.
	Issue 328: BEAUti: No longer generates very long default value and bound
	           arrays for parameters.
	Issue 332: BEAUti: Add traits now allows any trait name to be given.
	Issue 333: BEAUti: Unlink buttons are enabled even when only one partition is
	           loaded
	Issue 337: BEAUti: Creating a 'CONTINUOUS' or 'INTEGER' trait gives Discrete
	           model of substitution in SiteModelPanel
	Issue 345: BEAUti: traits can be edited in the Traits table.
	Issue 346: BEAUti: Show user a guide for an example of *BEAST or trait mapping
	           file format.
	Issue 347: BEAUti: Change for BEAUti default prior of ucld.stdv and kappa
	Issue 348: BEAST: For GUI versions, <close> then <cancel>, should allow BEAST
	           to continue running.
	Issue 349: BEAUti: Window is too large for low resolution projector
	           or screen.
	Issue 352: BEAUti: NullPointerException if leave any trait value
	           empty after activate BSSVS
	Issue 353: BEAUti: *BEAST is now explicitly activated by a check box
	           above the data partition table.
	Issue 356: BEAUti: Added instructions to show user how to load specified
			   starting tree from BEAUti
	Issue 357: BEAST: Linux/UNIX version was unable to locate the native libraries.
	Issue 361: BEAUti: RegEx for guess dates and guess traits now works better.
	Issue 369: Beauti: Tip Dates panel: Tip date sampling "apply to" resets to
			   "All taxa" after navigating to different tab
	Issue 374: BEAUti: only one taxon set for 3 unlinked partitions
	Issue 380: BEAUti: Discrete trait model should have its tree linked to that
			   for a tree for a sequence partition
	Issue 381: BEAUti: Discrete trait model can select an inappropriate substitution
			   model in the data partition table
	Issue 383: BEAUti: "species" trait should not appear in Data Panel
	Issue 384: BEAST: Unhandled attribute (typo?) name in sumStatistic:rateChanges
	Issue 387: BEAUti: Generates wrong xml when a uniform prior is selected
			   for species.popMean
	Issue 391: BEAUti: add includeStem for taxon set

Version 1.5.4 released 29th March 2010

Bug Fixes
  • This version addresses the following issues: 
        Issue 52:  Commas put into years displayed in BEAUti
	Issue 59:  examples/testCoalescent.xml
	Issue 74:  mcmc.operators
	Issue 161: BEAUTi: review all initial rates and root heights
	           calculations
	Issue 162: Fix unstable JUnit test
	Issue 198: Improper JUnit test code causes incorrect failure in
	           Hudson
	Issue 241: BEAUTi: operator subtreeSlide size should not be INF
	Issue 270: BEASTParserDoc is unable to generate Wiki format
	Issue 267: BEAUTi: The user should *always* be allowed to fix
	           the rates if they want
	Issue 268: BEAUti: prior distribution error message "find
	           incorrect number format"
	Issue 277: Typo in BEAUTi - Unable to overwrite exsting file
	Issue 282: BEAUti: AlignmentGenerator cannot print sequence with
	           long sites properly
	Issue 285: BEAST: ComplexSubstitutionModelParser can only refer to
	           SVS not ComplexSubstitutionModel
	Issue 286: BEAUti: remove exponential in VariableDemographicModel
	           from BEAUti
	Issue 287: BEAUti: EBSP VariableDemographicModel popSize value =
	           populationMean value and indicators value = 0
	Issue 288: BEAUti: avoid comma generated in the numbers in the log

        All issues can be viewed at http://code.google.com/p/beast-mcmc/issues/list

Version 1.5.3 released 9th December 2009

New Features
  • GUI versions of BEAST show a dialog box to set various options that are otherwise only available by command-line.
Bug Fixes
  • This version addresses the following issues: 
        Issue 55:  MCMC Screen log doesn't currently work.
        Issue 70:  beauty and partition into codon positions
        Issue 72:  Beauti should hard limit GTR model parameters
        Issue 122: Ctrl-A in Beauti -> Exception
        Issue 127: BEAUTi prior panel, prior upper and lower does not show
                   INFINITY
        Issue 128: BEAUTi: tree prior Logistic Growth is not available for
                   calibration
        Issue 146: BEAUTi: binary COVARION model
        Issue 169: Editing the file name stem in BEAUti causes the insertion
                   point to jump to the end
        Issue 173: BEAUTi: partially link exception
        Issue 181: BEAUTi: need enhancement to default behaviour when adding
                   additional alignment after unlinking all
        Issue 193: Improved convergence diagnostic: compute effective size of
                   combined runs by concatenating the samples
        Issue 194: BEAUTi: prior setting exception when changing value
        Issue 199: Tracer shows only first log file entry when multiples
                   entries have the same name
        Issue 201: Editing file name stem in MCMC panel causes insertion point
                   to jump to end
        Issue 202: Build scripts producing JDK 1.6 classes so distribution fails
                   on machines with JDK 1.5
        Issue 203: Launch4j bug affecting the link of NucleotideLikelihoodCore.dll
        Issue 204: BEAUTi: tree prior Constant Size is not working for node calibration
        Issue 205: Tittle information disappeared in the console using windows
                   paramter to run BEAST
        Issue 209: pathogen dose not compile
        Issue 211: BEAUTi: Amino Acid site model xml is wrong
        Issue 212: BEAST Linux version: need chmod 755 *
        Issue 215: Default memory allocations for Mac packaging are too small
        Issue 216: Default memory allocations for Linux packaging are too small
        Issue 217: Problems with the Mac OS X packaging
        Issue 218: BeastMain options dialog giving error "Illegal entry value
                   must be between 1 and 21474863647"
        Issue 220: The shell scripts to launch BEAST in command line can't cope
                   with spaces in the path.
        Issue 221: Reweight the substitution model operators in BEAUti
        Issue 222: Naming of 'treeLikelihood' when only 1 partition
        Issue 223: Set the 'operator analysis' file name as an attribute in
                   MCMC element
        Issue 225: TreePartitionCoalescent is not existing
        Issue 226: The parsers in release_parsers.properties need to be updated
        Issue 227: Fix BeastDoc to generate tag doc for BEAST users
        Issue 229: Another failure of testing framework
        Issue 231: Delay the start of the performance measurement until after
                   full evaluation
        Issue 232: Mac (Windows?) version of BEAST closes console on error
        Issue 234: Make the dreaded "The initial posterior is zero" error a bit
                   more friendly to debug
        Issue 236: dr.evomodel.operators.TraitGibbsOperator is duplicated with
                   parsers in either release_parser.properties or BeastParser.java
        Issue 237: BEAUTi: taxon set not handling multi-tree partition
        Issue 238: Naming of 'patterns' when only 1 partition
        Issue 240: BEAUTi: useAmbiguities of treeLikelihood should be default
                   as true for binay Covarion
        Issue 242: BEAUTi: create a check box to choose useAmbiguities in
                   treeLikelihood for binary data
        Issue 244: URGENT: Default prior review
        Issue 247: BEAUTi: wrong XML in frequecies choosing EMPIRICAL binary
                   COVARION model
        Issue 248: BEAUTi exception (*BEAST) when linking some same taxa for a
                   tree model by given different taxa in total
        Issue 250: BEAUTi: wrong alignment blocks for multi-gene when choosing
                   "create empty alignment"
        Issue 253: BEAUTi add a pop-up box to check the default priors when
                   click Generate BEAST file
        Issue 254: BEAUTi generate XML button should be disable after all data
                   partition removed
        Issue 255: Turn off parser WARNINGs by default for released version
        Issue 256: BEAUTi: Add comment in XML "skyride.logPopSize is log unit
                   unlike other popSize"
        Issue 257: BEAUTi: Binary data and multiple partitions does not generate
                   the siteModel correctly
        Issue 259: BEAUTi: Mac OS MCMC panel problem
        Issue 260: BEAUTi: correct English
        Issue 261: Error parsing <traceAnalysis> element with id null
        Issue 262: wrong XML of testCataclysmCoalescent.xml
        Issue 264: ImportException: Number format error for <treeTraceAnalysis ...

        All issues can be viewed at http://code.google.com/p/beast-mcmc/issues/list

Version 1.5.2 released 9th October 2009

New Features
  • GUI versions of BEAST show a dialog box to set various options that are otherwise only available by command-line.
Bug Fixes
  • Many bugs fixed that were introduced in 1.5.1 including: 
        Issue 44: Recent change (svn commit 773) breaks Beauti
        Issue 75: "Unlink substitution model across codon positions has no effect
        Issue 77: BEAUti v1.5 produces decimal operator weights that don't work with
                  old BEAST
        Issue 94: Improve the prefix of PartitionClockModelTreeModelLink
        Issue 123: BEAUti uses a poor choice of name for re-linking partitions  
        Issue 131: BEAUTi: rewrite Parameter and Operator class by Builder pattern,
                   and all referring methods
        Issue 134: TN93 parser problem
        Issue 135: Laplace prior has not been implemented in PriorParsers, but
                   appearing in BEAUTi.
        Issue 136: BEAUTi log screen should show actual name instead of Rate 1, 2, ....
        Issue 139: ImportnaceNarrowExchange should be ImportanceNarrowExchange
        Issue 140: more options for *beast
        Issue 144: BEAUTi: "In use" in Operators panel cannot be changed for those
                   operators belonging to PartitionClockModelTreeModelLink  
        Issue 145: BEAUTi: implement frequencies for binary model in the same way as
                   for DNA data
        Issue 147: BEAUTi: if no leaf nodes selected in Taxon Set panel, the
                   corresponding tmrca should not appear in Prior panel
        Issue 148: BEAUTi: illegal entry error for Clock Model panel
        Issue 151: BEAUti: MCMC tab numerical fields reset after GUI events
        Issue 154: No Nucleotide Likelihood Core in java.library.path
        Issue 155: Warning message " ... does not match stored object with same id
                   and tag name ..." confused user
        Issue 156: LogNormal in prior panel in Beauti
        Issue 157: LogNormal in Beauti: Exception occurred during event dispatching
        Issue 158: We should have the Inverse Gamma distribution in Beauti
        Issue 160: Ant build for Mac version release
        Issue 163: BEAUTi: unlink trees if allow different taxa
        Issue 164: TreeAnnotator v1.5.1 incorrectly labels tree [bug report AND fix
                   solution!]
        Issue 165: BEAUTi: some priors parameters keep changing
        Issue 168: GMRF 'Skyride' complains about 'Share same tree prior' even if only
                   one tree
        Issue 172: BEAUti: shareSameTreePrior should be disabled when there is only 1
                   tree model
        Issue 176: BEAUti: LogNormal mean should be > 0 when mean in real space
        Issue 179: species.popSize is a horrible choice of name
        Issue 180: BEAUTi: allMus is not dealing with multi-partition + codon, and
                   allClockRate
        Issue 183: BEAUTi: clock model should use different mu with codon model
        Issue 184: BEAUTi: make "mu" only appeared using codon model
        Issue 185: Guess Trait dialog doesn't offer user defined traits
        Issue 188: Suggestion to remove 'pre-burnin'
        Issue 190: Mac app ant build needs several corrections
        Issue 192: BEAUti: GUI exception when attempting to add a new trait 
        Issue 195: BEAUTi ignores and resets change of prior
        Issue 196: Beast failure with file generated by BEAUTi (ucld)
        Issue 197: meanRate, coefficientOfVariation and covariance should have gene name
                   in name="" attribute
        Issue 200: BEAUTi: *BEAST using Import Trait cannot add species in Trait table

        All issues can be viewed at http://code.google.com/p/beast-mcmc/issues/list

Version 1.5.1 released 3rd September 2009

New Features
  • *BEAST: a version of BEAST that can estimate species trees from multi-locus multi-individual data sets using the multispecies coalescent (Heled & Drummond, 2009, submitted)
  • BEAUti overhaul: BEAUti has been completely re-written to allow flexible handling of multi-partition analyses, including partitioning by substitution models, clock models and trees. BEAUti also supports the new *BEAST method.
  • Relaxed clock refinements: The uncorrelated relaxed clock models have been refined by the addition of two new operators to address (valid) criticisms levelled in "Bayesian estimation of species divergence times from multiple loci using multiple calibrations", Rannala and Yang (2007). The change allows more than one branch to have the same relaxed clock rate category. In practice this will only have any appreciable impact on trees that have a small number of branches (< 15 taxa).
  • Added phylogeography models of Lemey et al (2009), PLoS Comput Biol.
  • Added post mortem damage models of Rambaut et al (2009) Mol Biol Evol.
Bug Fixes
  • Many bugs fixed that were introduced in 1.5.0 including: 
		
		Issue 38: If partitions are unlinked then the base frequencies should be 
	              calculated from each partition separately;
		Issue 80: ant build generates a broken beast;
		Issue 81: ant build release fails after recent changes to beauti;
		Issue 82: dr.app.util.Utils is broken (use of java 1.6 features?);
		Issue 97: Remove "Link All/Unlink All" buttons from Data Tab;
		Issue 98: Request to be able to view data partition alignments;
		Issue 99: Suggestion: Clock Models tab;
		Issue 100: Request to revise UI of "Choose Fix Rate Option";
		Issue 103: Comments and suggestions for "MCMC Tab";
		Issue 106: Single locus but multiple codon partitions: Doesn't allow 
		           'Fixed mean rate';
		Issue 109: BEAUti: Generating XML - no file overwriting warning?;
		Issue 110: BEAUTi: excludedTaxonSetsComboBox and includedTaxonSetsComboBox 
		           in Taxon Sets panel not updating
		Issue 114: Beauti Priors panel for starbeast;
		Issue 115: Beauti starbeast: missing updown operator for first gene;
		Issue 116: Make TN93 available in BEAUTi;
		Issue 118: Beauti defaults - clock.rate in starbeast;
		Issue 119: Beauti fails;
		Issue 120: Beauti silliness - warn about not saving when nothing was done;
		Issue 121: Beauti generates spurios multiple upDown operators;
		Issue 124: Create method to check calibration situation;
		Issue 125: Tree prior panel affecting isShareSameTreePrior check box wrongly.
		        
		All issues can be viewed at http://code.google.com/p/beast-mcmc/issues/list

Version 1.5.0 released 4th August 2009

This was a major new version with numerous new features and many bugs - see release 1.5.1 for details.
Bug Fixes
  • Many new bugs were introduced.

Older Versions

Version 1.4.8 released 27nd July 2008

New Features
  • A birth-death model of speciation. This is the model of Tanja Gernhard (The conditioned reconstructed process, J Theor Biol, 2008).
  • There is now an option to dynamically rescale the likelihoods of trees to avoid loss of precision. This can avoid problems with large trees but may affect performance. See on-line documentation for details.
Bug Fixes
  • Improvements to how uncorrelated relaxed clocks are sampled. This will be an improvement more for small trees than large.
  • Fixed a problem setting the number of groups for the Bayesian Skyline Plot in BEAUti.
  • A warning message is given if a speciation process is selected for serially-sampled sequences.
  • LogCombiner ignores lines starting with a '#' as a comment line.
  • Trees are written with a standard (alphabetical) ordering of taxa to allow them to be read by AWTY.
  • For writing NEXUS tree files, taxon names with illegal characters (such as spaces or semicolons) are now quoted.
  • LogCombiner now uses the '.' as the decimal separator irrespective of country or language. This is to standardize with other programs.

Version 1.4.7 released 1st February 2008

New Features
  • BEAUti can now set nucleotide frequencies to be sampled or equal as well as the empirically estimated (the previous default).
  • Individual operators can now be turned on and off in BEAUti (for example to sample on a fixed tree topology).
Bug Fixes
  • BEAST now generates starting trees with certain taxon constraints that caused problems before.
  • Some numerical problems with the Expansion and Logistic models have been fixed or at least improved for certain situations.
  • Fixed a problem in which BEAUti's Prior Distribution dialog was freezing when a very large lognormal mean was set.
  • A seed of zero crashed BEAST - seeds must be > 0 now.
  • The command-line version of TreeAnnotator wasn't accepting a target tree file name.
  • Fixed a problem in LogCombiner where if the Run button was pressed before the user had finished editing a burnin the value was ignored.
  • Fixed an issue with BEAUti running under Mac OS X 10.5 (Leopard) where the text box for the substitution rate was tiny.

Version 1.4.6 released 26th September 2007

Bug Fixes
  • Fixed a problem in which the lognormal distribution wasn't giving the required distribution as prior.
  • Fixed a problem with TreeAnnotator not completing when finding the Maximum Clade Credibility tree.

Version 1.4.5 released 15th August 2007

New Features
  • BEAUti has ability to generate BEAST files that sample only from the prior distribution by replacing sequences with 'N's. This option can be found in the MCMC panel.
Bug Fixes
  • Defining taxa sets that weren't compatible could cause BEAST to fail to run.
  • The priors table didn't update when the tree prior was changed.
  • The parametric plots in the Prior Dialog now show correctly.
  • Introduced a default Jeffery's prior on certain parameters (not technically a bug fix but may make the defaults more robust).

Version 1.4.4 released 12th July 2007

New Features
  • Improved speed of relaxed clock by more judicious updating of the likelihood when rates change on a few branches.
  • Slight improvement in speed calculating the coalescent prior.
  • The taxa and translate lists in NEXUS trees files can now be sorted alphabetically (and this is the default in BEAUti). This will help comparisons of trees files using certain software that assume taxon orders are the same (namely AWTY).
Bug Fixes
  • Log normal prior element required meanInRealSpace attribute which prevented older XML file from running without modification. This attribute is now optional.

Version 1.4.3 released 18th June 2007

Bug Fixes
  • The Yule tree prior was wrongly implemented.
  • BEAUti did not allow some types of compatable monophyly constraints.
  • BEAST did not always simulate starting trees with enforced monophyly correctly.
  • BEAST ignored specified offsets for exponential priors on parameters.
  • BEAUti ignored user specified offsets for gamma distribution priors on parameters.

Version 1.4.2 released 25th April 2007

New Features
  • Taxon sets can now have monophyly enforced using BEAUti.
  • Starting tree is now compatible with monophyly constraints and uniform priors on tMRCAs.
  • Relaxed clock statistics (meanRate, etc.) are now in the priors table.
Bug Fixes

Version 1.4.1 released 25th January 2007

Bug Fixes
  • BEAUti did not accept the number of skyline groups properly.
  • BEAUti was not able to 'guess dates' using a specified prefix.
  • Fixed a problem using exponential distribution as priors.
  • LogCombiner could mix up taxon labels when combining multiple tree files. This problem would not affect the reconstruction of skyline plots.

Version 1.4 released 31th October 2006

New Features
  • Faster likelihood calculations after rejecting a state. Whilst the improvement in speed will vary, the indications are that this is a significant overall speed improvement.
  • Rates for relaxed clock models can be written embedded in the tree for later analysis.
  • 2 new accessory programs: LogCombiner for combining multiple runs into single log files and TreeAnnotator which annotates a given target tree with posterior estimates from a BEAST sample of trees.
  • BEAUti can now create TaxonSets and apply prior probability distributions to the tMRCA of these.
  • BEAUti can now save and load "Template Files" containing model settings but not data.
  • BEAUti can be run non-interactively as a command-line using the template files.
  • The order of parameters logged in BEAST files created by BEAUti has been changed to make it easier to compare files with different models in Tracer.
Bug Fixes
  • BEAUti did not use the specified fixed substitution rate.
  • Fixed a problem using UPGMA as an initial tree - would sometimes give ? in log files instead of posteriors.
  • Fixed a problem with BEAUti not loading data when using a non-english operating system.
  • Fixed a problem using gamma distribution as priors.

Version 1.3 released 30th August 2005

New Features
  • "Relaxed Clock" models.
  • New prior selection for parameters using BEAUti.
Bug Fixes
  • Problem with distributionLikelihood including all data twice resulting in the distributions being too tight.
  • Improved error reporting.

Version 1.2 released 24th February 2005

New Features
  • "Bayesian Skyline" demographic model.
  • Logistic demographic model.

Version 1.1.2 released 27th August 2004

Bug Fixes
  • Amino Acid models were not being set correctly
  • resulting in the use of an equal-rates model.

Version 1.1.1 released 24th August 2004

New Features
  • Can now switch autoOptimize on or off for individual operators. This overrides the setting for the <mcmc> element.
  • Default weights for certain operators have been changed in BEAUti to improve performance.
Bug Fixes
  • Fixed a bug in BEAUti in which unlinked codon analyses produced non-working XML files.
  • Fixed a bug where BEAST wouldn't run certain old XML files throwing a null-pointer exception when writing a log file.
  • Doesn't stop with RuntimeException when a state is proposed that has a really bad (underflowing) likelihood.

Version 1.1 released 5th August 2004

New Features
  • An auto-optimize function that removes the need to perform manual tuning of operators.
  • Improvement in speed of likelihood calculations for nucleotides.
  • Now includes BEAUTi as part of the BEAST package. In future these will be distributed together with matching version numbers.
Bug Fixes
  • Various minor problems with parsing XML have been fixed

Version 1.0.3 released 12 July 2003

New Features
  • Improved tree log analyser. Now gives mean node heights (branch lengths) on trees with a frequency > 100.
Bug Fixes
  • Various minor problems with parsing XML have been fixed

Version 1.0.2 released 27 June 2003

New Features
  • Improved tree log analyser. Now gives mean node heights (branch lengths) on trees with a frequency > 100.
Bug Fixes
  • Fixed the wrong coalescent likelihood function introduced in the last version.
  • Various minor problems with parsing XML have been fixed (including GTR model, NormalDistribution and others).

Version 1.0.1 released 23 June 2003

This version introduced a 'fix' for the coalescent likelihood function which turned out to be erroneous.
This version should not be used.

Version 1.0 released 10 June 2003

First released version.
Return to contents

Alexei Drummond, Andrew Rambaut and Marc Suchard
Copyright © 2002-2010 All rights reserved.

Retrieved from "http://beast.bio.ed.ac.uk/Version_History"
Personal tools