Phylemon: a suite of web tools for molecular evolution, phylogenetics and phylogenomics

Joaqun T arraga 0 1 Ignacio Medina 1 Leonardo Arbiza 1 Jaime Huerta-Cepas 0 1 Toni Gabaldo n 1 Joaqun Dopazo 0 1 Herna n Dopazo 1 0 Functional Genomics Node , INB, CIPF, Valencia 46013, Spain 1 Bioinformatics Department, Centro de Investigacio n Prncipe Felipe (CIPF) Phylemon is an online platform for phylogenetic and evolutionary analyses of molecular sequence data. It has been developed as a web server that integrates a suite of different tools selected among the most popular stand-alone programs in phylogenetic and evolutionary analysis. It has been conceived as a natural response to the increasing demand of data analysis of many experimental scientists wishing to add a molecular evolution and phylogenetics insight into their research. Tools included in Phylemon cover a wide yet selected range of programs: from the most basic for multiple sequence alignment to elaborate statistical methods of phylogenetic reconstruction including methods for evolutionary rates analyses and molecular adaptation. Phylemon has several features that differentiates it from other resources: (i) It offers an integrated environment that enables the direct concatenation of evolutionary analyses, the storage of results and handles required data format conversions, (ii) Once an outfile is produced, Phylemon suggests the next possible analyses, thus guiding the user and facilitating the integration of multi-step analyses, and (iii) users can define and save complete pipelines for specific phylogenetic analysis to be automatically used on many genes in subsequent sessions or multiple genes in a single session (phylogenomics). The Phylemon web server is available at http://phylemon.bioinfo.cipf.es. - Phylogenetic and evolutionary analyses of sequences are among the most often used methodologies in laboratories working in functional, comparative and structural genomics (1). Since 1980, when the first version of the PHYLogeny Inference Package (PHYLIP) (2) was introduced by Felsenstein, a high number of programs for phylogenetic inference have been developed. Currently, PHYLIP(2), PAUP (3), MEGA(1), PhyML(4), PAML(5) and MrBayes(6) are well-known programs that are used by thousands of users around the world. Other more specific programs, designed to test evolutionary hypotheses for model selection, tree topology, molecular clock or adaptation, are less popular among common users, but they are, nevertheless, of great interest for users familiar with evolutionary enquiries. Currently, the most comprehensive list of phylogenetic resources can be found at the University of Washington in Seattle (http://evolution. genetics.washington.edu/phylip/software.html), which listed 292 phylogeny packages and 38 web servers, by 2003. Web servers for phylogenetic and evolutionary analyses provide a direct means for addressing several evolutionary questions, ranging from the computation of a multiple alignment and a neighbor-joining tree using ClustalW program (7) (http://www.ebi.ac.uk/clustalw/), to the more sophisticate analysis of molecular adaptation for detection of positively selected sites in DNA sequences (using methods as those available in the HYPHY package (8) (http://www.datamonkey.org/)). Many such servers run a single tool or program whereas others bring together many of the most popular programs of phylogenetic reconstruction (e.g. see http://bioweb.pasteur.fr/seqanal/ phylogeny/intro-uk.html). Despite this diversity there is, so far, no single integrated web server that provides a common framework to run the most frequent analyses on DNA and protein sequences from a phylogenetic and evolutionary perspective. Non-expert users are then often overwhelmed by the variety of servers, formats and options available and by the difficulty of concatenating analyses performed on different servers. The main objective of Phylemon is to fulfil this need by providing users with the possibility of finding all the necessary applications in a single integrated web framework that guides them throughout the whole evolutionary analysis. OUTLINE OF THE PROGRAM Phylemon is a web server that integrates a selected suite of more than 20 different tools from the most popular standalone programs of phylogenetic and evolutionary analysis (Figure 1A). Three features characterize all tools integrated in Phylemon: (1) tools have available examples in order to familiarize users with the correct input data and expected results, (2) input formats (preferentially FASTA or PHYLIP) are automatically transformed in order to move among alternative tools and (3) all the input and output result files can be saved in default or user-defined projects (folders). Phylemon can be accessed by anonymous login or by registered users. The only difference between these choices is that registered users, from whom only an e-mail is required, can store project results and use them at a later time for further analysis (Figure 1G). PHYLOGENETIC PROGRAMS Phylemon runs distance-based methods, maximum parsimony analyses and statistical methods of phylogenetic reconstruction. Distances and parsimony methods for DNA or protein sequence data are provided by the most often used algorithms of the PHYLIP package (2) v3.65: DnaDist, ProtDist, DnaPars and ProtPars, respectively. Pairwise distance matrices can be represented in a phylogenetic tree using the neighbor-joining (NJ) algorithm (Neighbor) or applying a least square (LS) method or a minimum evolution (ME) criterium (Fitch program). In order to obtain trees with statistical support on internal nodes, a re-sampling method (i.e. bootstrap option included in the Seqboot algorithm) and the corresponding trees summarizing algorithm (i.e. majority rule tree using Consense program) of PHYLIP are included. Basic maximum likelihood (ML) analyses of DNA and protein sequence data are provided with the DnaML and ProML algorithms of the PHYLIP package in Phylemon. When a more sophisticated ML analysis is required users can run PhyML version aLRT (9,10) or TREE-PUZZLE v5.2 (11). Major differences between these ML programs are: (1) PhyML is faster than any other ML algorithm of phylogenetic reconstruction, (2) TREE-PUZZLE uses a quartet-puzzling method instead the more classical heuristic searches for tree searching, (3) TREE-PUZZLE reports reliability values while the PhyML method reports Felsensteins bootstrap values and aLRT-related statistics branch support (9), (4) TREE-PUZZLE can quantify the amount of the phylogenetic signal contained in a data set (the probability of the data producing a tree-like phylogenetic representation) through the likelihoodmapping method (12) and (5) TREE-PUZZLE computes ML pairwise distances that can easily be represented in an NJ/LS/ME tree. Finally, Phylemon runs Bayesian phylogenetic analysis using MrBayes v3.1.12 (6). MrBayes runs in Phylemon with the same characteristics that users have in Windows or Linux interfaces. Users can define all the parameters of MrBayes in a fil (...truncated)