- Open Access
TreeGraph 2: Combining and visualizing evidence from different phylogenetic analyses
© Stöver and Müller; licensee BioMed Central Ltd. 2010
- Received: 17 July 2009
- Accepted: 5 January 2010
- Published: 5 January 2010
Today it is common to apply multiple potentially conflicting data sources to a given phylogenetic problem. At the same time, several different inference techniques are routinely employed instead of relying on just one. In view of both trends it is becoming increasingly important to be able to efficiently compare different sets of statistical values supporting (or conflicting with) the nodes of a given tree topology, and merging this into a meaningful representation. A tree editor supporting this should also allow for flexible editing operations and be able to produce ready-to-publish figures.
We developed TreeGraph 2, a GUI-based graphical editor for phylogenetic trees (available from http://treegraph.bioinfweb.info). It allows automatically combining information from different phylogenetic analyses of a given dataset (or from different subsets of the dataset), and helps to identify and graphically present incongruences. The program features versatile editing and formatting options, such as automatically setting line widths or colors according to the value of any of the unlimited number of variables that can be assigned to each node or branch. These node/branch data can be imported from spread sheets or other trees, be calculated from each other by specified mathematical expressions, filtered, copied from and to other internal variables, be kept invisible or set visible and then be freely formatted (individually or across the whole tree). Beyond typical editing operations such as tree rerooting and ladderizing or moving and collapsing of nodes, whole clades can be copied from other files and be inserted (along with all node/branch data and legends), but can also be manually added and, thus, whole trees can quickly be manually constructed de novo. TreeGraph 2 outputs various graphic formats such as SVG, PDF, or PNG, useful for tree figures in both publications and presentations.
TreeGraph 2 is a user-friendly, fully documented application to produce ready-to-publish trees. It can display any number of annotations in several ways, and permits easily importing and combining them. Additionally, a great number of editing- and formatting-operations is available.
- Tree Figure
- Unlimited Number
- Editing Operation
- Graphical Editor
- Textual Annotation
It has become standard to apply multiple inference techniques to a given phylogenetic problem. The recent invasion of phylogenetics by Bayesian techniques (e.g., ), the ever improving models and algorithms for tree searches under maximum likelihood (e.g., [2, 3]), and the continuously growing processor speed helped these previously computationally very expensive approaches to become a typical component of most phylogenetic studies, accompanying the widespread parsimony and distance-based approaches. At the same time, no single inference technique has consistently proven to be the single best choice. Accordingly, the researcher is well-advised to explore potential method-specific differential results, leaving him or her with the difficulty of visualizing these differences for him- or herself and for the reader. Frequently, differences are restricted to the magnitude of various measures of statistical support (such as jackknife and bootstrap proportions, Bayesian posterior probabilities), rather than being apparent from the topology. In addition, the frequently reported results from topological tests (e.g., ) or tracing of ancestral character states (e.g., ) add further importance to being able to assign a variety of numbers and graphical labels to tree nodes.
To address those needs, the first version of TreeGraph  had been developed, which strongly simplifies the creation of the final tree figure by the automatic positioning and formatting of multiple labels per branch. However, while one support type could directly be imported from the phylogeny inference program output, the Newick- and Nexus  format used by these programs precluded the direct import of more branch labels. For all additional labels (support values), the laborious work of mapping them onto the appropriate nodes remained. The cumbersome drawing part of the publication process was minimized, but it remained the user's responsibility to collect and position all information that was to be displayed at the nodes.
We figured that automating this process would be very useful, particularly so in studies of extensive gene family datasets that may contain several hundred terminals. Gene family studies using phylogenetic approaches have become a major focus with the increasing amount of available fully sequenced genomes. Typically, gene family trees suffer from weak support [8–10]. The entailed caution required when interpreting gene family trees increases the need for testing alternative inference methods, alignment methods, data partitions, and varying treatment of questionable alignment regions.
Similarly, the differential contribution of and potential conflict among different data partitions is frequently estimated by the differential success of resolution and degree of statistical support in various parts of the tree contributed by each partition . This has become particularly important since multigene analysis are the rule rather than the exception, a trend further fueled by the growing availability of complete (organellar) genomes that provide easy access to a large number of genes that can be concatenated in large data matrices and then subjected to phylogenetic analyses, e.g. .
These trends call for a tree editor that is able to compare and ultimately visualize congruent and conflicting evidence from different analyses, while guaranteeing flexible editing and production of high-quality tree figures for publications.
TreeGraph 2 is written in Java and uses Swing for its graphical user interface (GUI) as well as the Apache Batik SVG Toolkit (http://xmlgraphics.apache.org/batik/), FreeHEP (http://java.freehep.org/), Java Math Expression Parser (http://sourceforge.net/projects/jep/) and BrowserLauncher (http://browserlaunch2.sourceforge.net/) libraries. Besides its GUI, which makes editing and formatting very intuitive, the current version 2 adds many features previously unavailable in the command line precursor and introduces an XML-based native file format (XTG).
TreeGraph 2 can read trees in Newick or Nexus format (including additional annotations in comments specified by BEAST ) as well as phyloXML tree descriptions  and can furthermore import annotations from text files generated e.g. with a spreadsheet application. Besides that, TreeGraph 2 facilitates combining information from different phylogenetic analyses of a given dataset. This is particularly useful e.g. in the study of extensive gene family datasets with large sets of terminals. The following sections describe this feature in greater detail.
Mapping statistical support onto congruent nodes
For each branch of a tree opened in TreeGraph 2, the corresponding support from other trees can be mapped whenever the topology defined by the current branch is present in them. Each of these other trees may represent the result from a different analytical approach or different data partition, and support values from these trees are assigned their own label ID by which they are grouped and amenable to future formatting or editing operations. Thus, all support values that stem from a particular analysis can be individually formatted e.g. by their relative position on the branch and/or their font and style.
Finding conflicting nodes and mapping contradictory support
In some studies not only the support from different analyses has been mapped onto the branches but also the strongest support for a contradictory topology was determined by inspection via eye [15, 16].
TreeGraph 2 uses the following algorithm automate this (for a better understanding it should be kept in mind that each branch splits a tree into exactly two subtrees).
Find the branch2 which defines a subtree subtree2 with the smallest number of terminals that contains all leafs of a subtree subtree1 defined by branch1.
Inside subtree2 find all branches that define a subtree which are on the one hand fully enclosed by subtree2 and on the other hand contain at least one terminal which is also part of subtree1 as well as at least one leaf which is not.
The highest support value in the set of these branches is added as a conflicting value onto branch1.
This highest conflicting support value can be distinguished from congruent values by user-specified formats, e.g. brackets, asterisks or different colors (see example in Figure 1).
Editing and formatting capabilities
The program features versatile editing and formatting options, such as automatically setting branch widths or colors according to the value of any of the unlimited number of variables that can be assigned to each node or branch.
Editing of node/branch data
Beyond typical editing operations such as tree rerooting and ladderizing or moving and collapsing of nodes, whole clades can be copied or cut out and placed into new empty files or inserted (along with all node/branch data) into other trees. Since nodes can also be manually added, whole trees can quickly be manually constructed starting from an empty file.
The editing operations are facilitated by versatile additive selection options that allow selecting many elements in a tree for subsequent formatting with just a few clicks. Additionally, every operation applied to an opened tree can be easily undone or redone using the undo-function.
Searching, replacing and translating tree leaf names
Searching and replacing is possible across all node/branch data columns (including taxon names and node labels).
More restrictive alignment file formats do not allow lengthy taxon names, so names get truncated. In other cases, the often clumsy taxon- or lab IDs used during a study survive up to the final alignment, phylogenetic dataset and the trees constructed from it until they need to be adjusted for the final tree to be presented in a paper. TreeGraph 2 can be requested to apply a translation table to use "cleaned" taxon names for the final output. This translation table can be constructed easily with help of the data export feature and any text editor or spread sheet program. Furthermore the lab IDs (old terminal names) can be saved in a hidden data field to be able to identify the terminals by these lab IDs so that additional support values could still be added later on.
Formatting document elements
Automatically setting line width, text height, and color
TreeGraph 2 allows automatically setting all formats (e.g. branch widths, branch colors, text colors, text heights, icon sizes) according to the value of a chosen node/branch data column. This provides a very intuitive way to graphically present the relative magnitude of, e.g., certain types of support or rates assigned to branches (see Figure 2 and 3 for examples).
Different view modes
All editing operations are facilitated by a very convenient way to zoom in and out, fitting the zoom to the window size, and a miniature overview (Figure 2) for navigating large trees.
When applicable (i.e., given that branch length information is provided), trees can be displayed as phylogram or chronogram (Figure 3), with multiple options for adjusting a scale bar (to indicate e.g. time spans in chronograms, rates in ratograms, or branch lengths in phylograms).
Exporting to graphic formats and printing
TreeGraph 2 outputs various vector and (anti-aliased) pixel graphic formats. Among these are SVG, PDF, or PNG, supporting transparent background where this applies. Using the graphic export function of TreeGraph 2, the most adequate graphic formats, resolutions, and image sizes for manuscripts, presentation slides, or web pages, respectively, can be specified.
An extensive, continuously updated online help system is available under http://treegraph.bioinfweb.info/Help and can also be accessed (in a context-dependent manner) from within the program. Additionally, several video tutorials are offered there to get started with TreeGraph 2 (see http://treegraph.bioinfweb.info/Help/wiki/Tutorial:Main_page).
Comparison to previous software
Besides importing additional annotations from tables (which TreeDyn also offers), TreeGraph 2 is the only editor which can combine annotations (e.g. statistical support from different analysis methods) from different trees (with the same set of terminals). The information gained this way has a topological component and can therefore not simply be obtained from data in a table.
A feature closely related to the ones mentioned above is the ability to calculate numeric or textual annotations by mathematical expressions which can reference other annotations (see above). To date, a similar functionality is not offered by any other tree editor.
TreeGraph 2 features a multitude of format options which can be combined to every tree element (e.g. branches, nodes or labels) independently. As Figure 4 shows, no other tree editor currently provides functionalities like element-specific formats for all types of tree elements in combination with advanced selection options or collision free positioning of the whole tree. Moreover, none of the editors that offer at least some of TreeGraph 2's formatting options allow the user to precisely determine the print layout. In contrast to most other editors, our program offers context help buttons (which link to the online help system) everywhere in the program, making it very easy for new users to get started.
It should be noted, however, that TreeGraph 2 has been optimized as a tree editor for producing high quality tree figures and not as a viewer for trees with many thousands of taxa which could never be depicted completely in a publication or presentation. The latter is a specialty of software specifically designed for this purpose such as, e.g., Dendroscope  (Figure 4).
Since TreeGraph 2 is written in Java and is able to read and write all its supported formats directly from and to streams in would be possible to use it in a web application either on the server (e.g. with Apache Tomcat) or the client site (e.g. as an Java applet or a Java webstart application) to display and manipulate trees. As yet, our application would have to be integrated into such a web application by its programmer manually and we do not yet offer a ready-to-use plug-in solution for this. We do, however, offer a full documentation of our source code (including its interfaces) to facilitate such a web integration.
With its easy-to-use graphical user interface and a number of semi-automatic editing and formatting options, TreeGraph 2 is a graphical editor useful in the context of any phylogenetic study. It is particularly useful where multiple, potentially conflicting trees are being produced, because its automatic combination of information from different analyses helps to identify and graphically present such incongruences. The way in which data can be imported and then assigned to nodes, manipulated or even converted to color tones, line diameters or other formats allows for a great flexibility in visualizing any kind of data associated with different parts of the tree. Together with the possibility to manually construct new clades or delete clades and the various graphic output formats supported, TreeGraph 2 greatly reduces the effort during the preparation of tree figures for presentations or publications.
Project name: TreeGraph 2
Project home page: http://treegraph.bioinfweb.info/ (including an extensive documentation and a development section with Javadocs)
Operating system(s): Platform independent (Java 6 has to be available)
Programming language: Java
Other requirements: Java Runtime Environment 6.0 (or higher)
License: GNU General Public License
Restrictions to use by non-academics: none
This work was in part supported by DFG grant MU2875/2 to KFM, since many of the features were added to the program in response to requirements encountered during work in the corresponding DFG project "Carnivory in Lamiales - Understanding character evolution, substitution rate plasticity, and genome miniaturization". Financial support to KFM by the Young Academy of the North Rhine-Westphalian Academy of Sciences (Nordrhein-Westfälische Akademie der Wissenschaften und der Künste) is highly appreciated. Thanks to Mark Simmons, Claude dePamphilis, Dietmar Quandt, and Jörn Müller for helpful suggestions. Finally we want to thank the authors of the open source libraries used.
- Holder M, Lewis PO: Phylogeny Estimation: Traditional and Bayesian Approaches. Nature Reviews Genetics 2003, 4: 275–284. 10.1038/nrg1044View ArticlePubMedGoogle Scholar
- Whelan S: Spatial and Temporal Heterogeneity in Nucleotide Sequence Evolution. Molecular Biology and Evolution 2008, 25: 1683–1694. 10.1093/molbev/msn119View ArticlePubMedGoogle Scholar
- Stamatakis A: RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics 2006, 22: 2688–2690. 10.1093/bioinformatics/btl446View ArticlePubMedGoogle Scholar
- Shimodaira H: An approximately unbiased test of phylogenetic tree selection. Systematic Biology 2002, 51: 492–508. 10.1080/10635150290069913View ArticlePubMedGoogle Scholar
- Pagel M, Meade A, Barker D: Bayesian Estimation of Ancestral Character States on Phylogenies. Systematic Biology 2004, 53: 673–684. 10.1080/10635150490522232View ArticlePubMedGoogle Scholar
- Müller J, Müller K: TreeGraph: automated drawing of complex tree figures using an extensible tree description format. Molecular Ecology Notes 2004, 4: 786–788. 10.1111/j.1471-8286.2004.00813.xView ArticleGoogle Scholar
- Maddison DR, Swofford DL, Maddison WP: Nexus: An extensible file format for systematic information. Systematic Biology 1997, 46: 590–621.View ArticlePubMedGoogle Scholar
- Barakat A, Müller KF, Sáenz de Miera LE: Molecular evolutionary analysis of the Arabidopsis L7 ribosomal protein gene family. Gene 2007, 403: 143–150. 10.1016/j.gene.2007.08.002View ArticlePubMedGoogle Scholar
- Sampedro J, Lee Y, Carey RE, dePamphilis C, Cosgrove DJ: Use of genomic history to improve phylogeny and understanding of births and deaths in a gene family. Plant Journal 2005, 44: 409–419. 10.1111/j.1365-313X.2005.02540.xView ArticlePubMedGoogle Scholar
- Zahn LM, Leebens-Mack JH, Arrington JM, Hu Y, Landherr LL, dePamphilis CW, Becker A, Theissen G, Ma H: Conservation and divergence in the AGAMOUS subfamily of MADS-box genes: evidence of independent sub- and neofunctionalization events. Evolution & Development 2006, 8: 30–45. 10.1111/j.1525-142X.2006.05073.xView ArticleGoogle Scholar
- Müller K, Borsch T, Hilu KW: Phylogenetic utility of rapidly evolving DNA at high taxonomical levels: contrasting matK, trnT-F and rbcL in basal angiosperms. Molecular Phylogenetics and Evolution 2006, 41: 99–117. 10.1016/j.ympev.2006.06.017View ArticlePubMedGoogle Scholar
- Jansen RK, Cai Z, Daniell H, Raubeson L, DePamphilis CW, Leebens-Mack J, Müller KF, Guisinger-Bellian M, Haberle RC, Hansen AK, et al.: Analysis of 81 Genes from 64 Chloroplast Genomes Resolves Relationships in Angiosperms and Identifies Genome-Scale Evolutionary Patterns. Proceedings of the National Academy of Sciences of the United States of America 2007, 104: 19369–19374. 10.1073/pnas.0709121104View ArticlePubMedPubMed CentralGoogle Scholar
- Drummond AJ, Rambaut A: BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evol Biol 2007, 7: 214. 10.1186/1471-2148-7-214View ArticlePubMedPubMed CentralGoogle Scholar
- Han MV, Zmasek CM: phyloXML: XML for evolutionary biology and comparative genomics. BMC Bioinformatics 2009, 10: 356. 10.1186/1471-2105-10-356View ArticlePubMedPubMed CentralGoogle Scholar
- Richardson D, Simmons M, Reddy A: Comprehensive comparative analysis of kinesins in photosynthetic eukaryotes. BMC Genomics 2006, 7: 18. 10.1186/1471-2164-7-18View ArticlePubMedPubMed CentralGoogle Scholar
- Zhang LB, Simmons MP: Phylogeny and delimitation of the Celastrales inferred from nuclear and plastid genes. Systematic Botany 2006, 31: 122–137. 10.1600/036364406775971778View ArticleGoogle Scholar
- Zmasek CM, Eddy SR: ATV: display and manipulation of annotated phylogenetic trees. Bioinformatics 2001, 17: 383–384. 10.1093/bioinformatics/17.4.383View ArticlePubMedGoogle Scholar
- Huson D, Richter D, Rausch C, Dezulian T, Franz M, Rupp R: Dendroscope: An interactive viewer for large phylogenetic trees. BMC Bioinformatics 2007, 8: 460. 10.1186/1471-2105-8-460View ArticlePubMedPubMed CentralGoogle Scholar
- Kumar S, Tamura K, Nei M: MEGA3: Integrated software for Molecular Evolutionary Genetics Analysis and sequence alignment. Briefings in Bioinformatics 2004, 5: 150–163. 10.1093/bib/5.2.150View ArticlePubMedGoogle Scholar
- Maddison WP, Maddison DR: Mesquite: A modular system for evolutionary analysis. Version 1.0. 2003.Google Scholar
- Jordan GE, Piel WH: PhyloWidget: web-based visualizations for the tree of life. Bioinformatics 2008, 24: 1641–1642. 10.1093/bioinformatics/btn235View ArticlePubMedGoogle Scholar
- Chevenet F, Brun C, Banuls A-L, Jacq B, Christen R: TreeDyn: towards dynamic graphics and annotations for analyses of trees. BMC Bioinformatics 2006, 7: 439. 10.1186/1471-2105-7-439View ArticlePubMedPubMed CentralGoogle Scholar
- Page RDM: TreeView: An application to display phylogenetic trees on personal computers. Computer Applications in the Biosciences 1996, 12: 357–358.PubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.