JCoDA: a tool for detecting evolutionary selection
© Steinway et al; licensee BioMed Central Ltd. 2010
Received: 12 January 2010
Accepted: 27 May 2010
Published: 27 May 2010
The incorporation of annotated sequence information from multiple related species in commonly used databases (Ensembl, Flybase, Saccharomyces Genome Database, Wormbase, etc.) has increased dramatically over the last few years. This influx of information has provided a considerable amount of raw material for evaluation of evolutionary relationships. To aid in the process, we have developed JCoDA (Java Codon Delimited Alignment) as a simple-to-use visualization tool for the detection of site specific and regional positive/negative evolutionary selection amongst homologous coding sequences.
JCoDA accepts user-inputted unaligned or pre-aligned coding sequences, performs a codon-delimited alignment using ClustalW, and determines the dN/dS calculations using PAML (Phylogenetic Analysis Using Maximum Likelihood, yn00 and codeml) in order to identify regions and sites under evolutionary selection. The JCoDA package includes a graphical interface for Phylip (Phylogeny Inference Package) to generate phylogenetic trees, manages formatting of all required file types, and streamlines passage of information between underlying programs. The raw data are output to user configurable graphs with sliding window options for straightforward visualization of pairwise or gene family comparisons. Additionally, codon-delimited alignments are output in a variety of common formats and all dN/dS calculations can be output in comma-separated value (CSV) format for downstream analysis. To illustrate the types of analyses that are facilitated by JCoDA, we have taken advantage of the well studied sex determination pathway in nematodes as well as the extensive sequence information available to identify genes under positive selection, examples of regional positive selection, and differences in selection based on the role of genes in the sex determination pathway.
JCoDA is a configurable, open source, user-friendly visualization tool for performing evolutionary analysis on homologous coding sequences. JCoDA can be used to rapidly screen for genes and regions of genes under selection using PAML. It can be freely downloaded at http://www.tcnj.edu/~nayaklab/jcoda.
The first step in the assessment of evolutionary relationships between related sequences is the generation of pairwise or multiple sequence alignments (MSAs). Over the last two decades several algorithms have been developed to generate rapid yet accurate sequence alignments for subsequent analysis . A commonly used program, ClustalW, generates MSAs of DNA or amino acids by constructing a branched guide tree from pairwise alignments . More recent progressive alignment methods, such as T-COFFEE, have improved the accuracy of ClustalW by combining information from local and global alignments . Other methods of sequence alignment, such as iterative, all-against-all, and hybrid approaches have also been shown to improve the accuracy of ClustalW, although some necessitate significant increases in computational power [4–7]. Regardless of the method, when DNA is aligned it is done so in a manner that arranges sequences to minimize gaps and mismatches to achieve a maximal score based on sequence identity and similarity.
There are currently several informatics tools freely available that calculate dN/dS ratios to measure evolutionary selection or generate codon-delimited alignments. The online program OCPAT is able to generate codon-delimited of alignments from human gene IDs and their putative orthologs from other vertebrates; however, OCPAT is not able to calculate dN/dS scores . Most of the programs that calculate dN/dS require a codon-delimited alignment. For example, the programs SNAP (Synonymous Non-synonymous Analysis Program) and WINA (Window Analysis) employ user provided alignments to calculate substitution rates, where WINA also allows for the use of sliding window analysis [10–12]. SWAKK (Sliding Window Analysis of Ka and Ks) uses pairwise sequence alignments, sliding windows, and structural alignment to identify regions of positive selection . DNaSP (DNA Sequence Polymorphism) allows for the detection of diversifying selection by measuring DNA polymorphisms and also uses sliding window analysis .
In an effort to simplify the process, PAL2NAL (v12) takes pre-aligned protein and the corresponding unaligned coding DNA to generate codon-delimited alignments. While useful, PAL2NAL is constrained to pairwise dN/dS analysis, does not include graphing, and does not allow the user to edit the codeml control file . The popular program PAML (Phylogenetic Analysis by Maximum Likelihood) is able to perform both pairwise and site-based dN/dS calculations. However, it does not have a GUI (graphical user interface) and visualization of the output data can be cumbersome. In addition, it requires the input of pre-aligned sequences . More recently, IDEA (Interactive Display For Evolutionary Analyses) has implemented a GUI for both PAML and Phylip (Phylogeny Inference Package) and is amenable for high throughput genome-wide analysis . While IDEA is a powerful tool, it does not run on the Windows operating system, requires the separate installation of several support programs, uses multiple languages, generates graphs that are difficult to configure, and also requires a codon-delimited alignment.
We have designed JCoDA to be an easy-to-use tool that integrates several common functions required for the detection of evolutionary selection in coding sequences. Specifically, the steps that are now controlled from the JCoDA interface: generation of codon-delimited alignments, generation of phylogenetic trees, estimation of site and regional dN/dS scores under multiple models of substitution, and the generation of user configurable graphical output. JCoDA only requires unaligned coding sequences (CDS) in FASTA format and it takes advantage of the freely available BioJava framework , ClustalW , Phylip , and PAML  to identify positive/negative evolutionary selection. The visualization options in JCoDA include a variety of common alignment formats, and user configurable scalable graphs with sliding window options for pairwise or gene family comparisons. To demonstrate the types of analyses facilitated by JCoDA, we have performed an analysis of 25 sex determination genes in nematodes and identified genes under positive selection, regional positive selection, and differences in selective pressure likely due to functional constraints.
JCoDA was designed to assist in performing common operations associated with evolutionary analysis. It coordinates the passage of information from the initial alignments (ClustalW version 2.0) to the calculation of dN/dS (yn00 or codeml, version 4.1) to the visualization of output. For site-based analysis of selection we have packaged JCoDA with a graphical interface for Phylip (version 3.6) that allows for the generation of phylogenetic trees. JCoDA is written entirely in Java, which allows for the addition of supplementary modules that offer additional functionality. To allow for easy installation, JCoDA is packaged with all of the required components (ClustalW, Phylip, and PAML) and can be installed on any computer with a Windows operating system (or virtual machine) and Java Runtime Environment 6 (JRE 6) (if Java 1.6 has been installed). For ease of operation, installation of Java Developer Kit 6 bundled with NetBeans (JDK, which includes JRE 6) is recommended http://java.sun.com/javase/downloads/index.jsp.
JCoDA provides two options for calculation of dN/dS scores: regional pairwise calculation via sliding window or a site-specific calculation. Pairwise dN/dS performs the calculation between any user-selected sequences from a list of all possible comparisons presented in the GUI. For sliding window pairwise calculations, the size of the window, jump, and substitution models are configurable via drop down menus. Once the selected sequence comparisons are submitted, JCoDA parses through them by window, converts them to Phylip format, and feeds each window to PAML (yn00) suite to calculate dN/dS. JCoDA does this iteratively over all windows until the end of the selected sequence pair. The potential benefit of the sliding window option is that it can be performed very quickly and it is able to extract information about regional selection. However, it is important to note that methods that incorporate sliding windows have been demonstrated to be prone to artifact arising due to resampling and illustrate the importance of incorporating site-based methods and the likelihood ratio test in sequence analysis .
JCoDA takes advantage of the codeml executable included in the PAML package to implement site-specific dN/dS calculation. Similar to pairwise comparisons, JCoDA converts all inputted sequences to Phylip format and feeds them to codeml. In order for a site-based dN/dS calculation to be performed, the user must provide a tree file and set its path for use in the GUI. To allow for this operation, the JCoDA package includes a Java based graphical interface for the Phylip package , http://evolution.genetics.washington.edu/phylip.html). The Phylip Graphical Interface (PGI) allows for the generation of trees using neighbor-joining, parsimony, or maximum likelihood based methods of either DNA or protein. Regardless of the source of the tree file, JCoDA will accept any tree in Phylip format.
JCoDA implements M7 (fit to a beta distribution, dN/dS > 1 disallowed) -vs- M8 (fit to a beta distribution, dN/dS > 1 allowed) and M1a (nearly neutral) -vs- M2a (positive selection) in conjunction with the likelihood ratio test to check for evidence of positive selection . To maintain JCoDA's flexibility, the user is given the option to access the complete codeml control file as a selectable advanced options tab to vary other parameters.
Results and Discussion
Input requirements and alignments
The JCoDA user interface, input requirements, and installation have been designed to be easier to use while retaining the underlying power of the codeml and yn00 programs from the PAML package (Additional File 1, Additional File 2, Additional File 3, Additional File 4 and Additional File 5). For example, we have simplified the input requirements to the coding sequences (CDS) in FASTA format. There were two primary reasons for this streamlined approach. First, CDS can be readily batch retrieved from NCBI or organism specific databases such as Wormbase (WormMart) . Second, the risk of mismatch between DNA and protein sequences is eliminated by directly translating the CDS input by the user.
For unaligned CDS, the sequences are translated by the BioJava framework  and the proteins are passed to ClustalW for alignment. JCoDA generates a codon-delimited alignment using the protein alignment from ClustalW as a guide to prevent the interruption of codon triplets, (Figure 1). JCoDA also accepts pre-aligned protein sequences in FASTA format paired with corresponding unaligned CDS in the same format to allow for the use of sequences aligned in another program or modified by hand. When provided with prealigned protein sequences and corresponding CDS, JCoDA simply circumvents ClustalW and performs the codon-delimited alignment directly.
Visualization and output
The use of JCoDA interface does not add significantly to the time need required to run ClustalW, Phylip, or PAML. Any time costs incurred are more than ameliorated by the integration of multiple tasks. For example, the user does not have to reformat sequences to shuttle them from one program to another or use an additional program to obtain graphical representations of the data. To test the ability of JCoDA to recover signatures of directional selection, we used multiple datasets including TRIM5α (hominids + OWMs) , nef gene of HIV-1 (pairwise and ML) , lysozyme , and a subset of HIV1 protease and reverse transcriptase sequences from the HIV positive selection mutation database . We have included an analysis of the sex determination genes in nematodes to illustrate the utility of JCoDA in identification of directional selection.
Evolutionary selection in the nematode sex determination pathway
Analysis of sex determination pathway genes
Proportion of sites with dN/dS>1
Relative to the genome wide C. elegans/C. briggsae dN/dS ratio of 0.06 , the majority of sex determination genes show elevated levels of non-synonymous substitutions in pairwise comparisons (Table 1, dN/dS, 0.06 -vs- 0.23). Considerable variability within each category was present but there are no significant differences in average pairwise dN/dS scores between the categories. Interestingly, site-based analysis reveals evidence for positive selection in five of nine genes involved exclusively in germ line sex determination and five of nine genes involved in the specification of somatic sex (dosage compensation). In contrast, positive selection was only detected in one gene of seven genes involved in both germ line and somatic pathways (Table 1, grey). Evidence of fewer genes that function in multiple pathways with signatures of positive selection likely reflects resistance to change based on additional function constraints.
Analysis of C. remanei fem-3 and tra-2 genes
Proportion of sites with dN/dS>1
Examples of the graphical output generated by JCoDA are shown in Figures 3 and 4. The GLD-1 (defective in germ line development) RNA binding protein governs the translation of numerous mRNA targets including functioning with FOG-2 in the promotion of spermatogenesis in the hermaphrodite [26, 28, 29]. As expected, pairwise sliding window analysis reveals that extensive purifying selection dominates regions with homology to the RNA binding GSG (GRP33/Sam68/GLD-1) domain (Figure 3, GSG domain). Interestingly, elevated levels of dN/dS were detected at both the N and C-terminal ends relative to the average pairwise dN/dS for gld-1 (0.07). Using site-based analysis, we confirmed the presence of residues under positive selection (Figure 4B). Curiously, even though GLD-1 orthologs in C. elegans and C. briggsae share significant amino acid homology (80% identity, 90% similarity) and have a dN/dS ratio consistent with purifying selection, they have opposite functions in sex determination . We speculate that species-specific functions could be explained by at least two mechanisms. First, the extensive conservation in the KH-domain between C. elegans and C. briggsae suggests that at least some of the species differences in GLD-1 function can be explained by a change in mRNA targets. Second, based on the elevated non-synonymous substations at the N and C-terminal ends, we can infer that some changes in GLD-1 function could result from species-specific interactions or regulation.
An important part of the analysis of homologous coding sequences is the characterization of evolutionary selection by comparing the rates of synonymous and non-synonymous substitutions. The primary issues with these types of analyses are the difficulty in generating codon-delimited alignments, shuttling between programs, and the complexity in configuring programs that are designed to detect positive selection. JCoDA provides a simple platform that integrates several common functions associated with evolutionary analysis of coding sequences and the detection of positive/negative selection. JCoDA is a modular tool built using the BioJava framework, ClustalW, Phylip, and PAML that allows for the rapid assessment and visualization of the pairwise and site-based selection pressure on coding sequences. Using JCoDA we were able to identify multiple sex determination pathway genes with evidence of positive selection based on functional constraints. The JCoDA executable, source code, and tutorial are freely available at http://www.tcnj.edu/~nayaklab/jcoda (Additional File 6).
Availability and requirements
Project name: JCoDA: A Tool for Detecting Evolutionary Selection
Project home page: http://www.tcnj.edu/~nayaklab/jcoda
Operating system(s): Windows
Programming language: Java
Other requirements: Java Runtime Environment 6.0 (JRE 6) or Java Developer Kit 6 (JDK 6, includes JRE)
JH and SN were funded in part by The College of New Jersey Mentored Undergraduate Summer Experience (TCNJ MUSE). We would like to thank Dr. Peter DePasquale (The College of New Jersey, Department of Computer Science) for helpful comments.
- Pirovano W, Heringa J: Multiple sequence alignment. Methods in molecular biology (Clifton, NJ) 2008, 452: 143–161. full_textView ArticleGoogle Scholar
- Thompson JD, Higgins DG, Gibson TJ: CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic acids research 1994, 22(22):4673–4680. 10.1093/nar/22.22.4673View ArticlePubMedPubMed CentralGoogle Scholar
- Notredame C, Higgins DG, Heringa J: T-Coffee: A novel method for fast and accurate multiple sequence alignment. Journal of molecular biology 2000, 302(1):205–217. 10.1006/jmbi.2000.4042View ArticlePubMedGoogle Scholar
- Egan A, Mahurkar A, Crabtree J, Badger JH, Carlton JM, Silva JC: IDEA: Interactive Display for Evolutionary Analyses. BMC bioinformatics 2008, 9: 524. 10.1186/1471-2105-9-524View ArticlePubMedPubMed CentralGoogle Scholar
- Gupta SK, Kececioglu JD, Schaffer AA: Improving the practical space and time efficiency of the shortest-paths approach to sum-of-pairs multiple sequence alignment. J Comput Biol 1995, 2(3):459–472. 10.1089/cmb.1995.2.459View ArticlePubMedGoogle Scholar
- Lipman DJ, Altschul SF, Kececioglu JD: A tool for multiple sequence alignment. Proceedings of the National Academy of Sciences of the United States of America 1989, 86(12):4412–4415. 10.1073/pnas.86.12.4412View ArticlePubMedPubMed CentralGoogle Scholar
- Sze SH, Lu Y, Yang Q: A polynomial time solvable formulation of multiple sequence alignment. J Comput Biol 2006, 13(2):309–319. 10.1089/cmb.2006.13.309View ArticlePubMedGoogle Scholar
- Yang Z: PAML 4: phylogenetic analysis by maximum likelihood. Molecular biology and evolution 2007, 24(8):1586–1591. 10.1093/molbev/msm088View ArticlePubMedGoogle Scholar
- Liu G, Uddin M, Islam M, Goodman M, Grossman LI, Romero R, Wildman DE: OCPAT: an online codon-preserved alignment tool for evolutionary genomic analysis of protein coding sequences. Source code for biology and medicine 2007, 2: 5. 10.1186/1751-0473-2-5View ArticlePubMedPubMed CentralGoogle Scholar
- Endo T, Ikeo K, Gojobori T: Large-scale search for genes on which positive selection may operate. Molecular biology and evolution 1996, 13(5):685–690.View ArticlePubMedGoogle Scholar
- Korber B: Computational Analysis of HIV Molecular Sequences. In HIV Signature and Sequence Variation Analysis. Edited by: Learn AGRaGH. Dordrecht, Netherlands: Kluwer Academic Publishers; 2000:55–72.Google Scholar
- Nei M, Gojobori T: Simple methods for estimating the numbers of synonymous and nonsynonymous nucleotide substitutions. Molecular biology and evolution 1986, 3(5):418–426.PubMedGoogle Scholar
- Liang H, Zhou W, Landweber LF: SWAKK: a web server for detecting positive selection in proteins using a sliding window substitution rate analysis. Nucleic acids research 2006, (34 Web Server):W382–384. 10.1093/nar/gkl272Google Scholar
- Rozas J, Rozas R: DnaSP version 3: an integrated program for molecular population genetics and molecular evolution analysis. Bioinformatics (Oxford, England) 1999, 15(2):174–175. 10.1093/bioinformatics/15.2.174View ArticleGoogle Scholar
- Suyama M, Torrents D, Bork P: PAL2NAL: robust conversion of protein sequence alignments into the corresponding codon alignments. Nucleic acids research 2006, (34 Web Server):W609–612. 10.1093/nar/gkl315Google Scholar
- Holland RC, Down TA, Pocock M, Prlic A, Huen D, James K, Foisy S, Drager A, Yates A, Heuer M, et al.: BioJava: an open-source framework for bioinformatics. Bioinformatics (Oxford, England) 2008, 24(18):2096–2097. 10.1093/bioinformatics/btn397View ArticlePubMed CentralGoogle Scholar
- Larkin MA, Blackshields G, Brown NP, Chenna R, McGettigan PA, McWilliam H, Valentin F, Wallace IM, Wilm A, Lopez R, et al.: Clustal W and Clustal X version 2.0. Bioinformatics (Oxford, England) 2007, 23(21):2947–2948. 10.1093/bioinformatics/btm404View ArticleGoogle Scholar
- Felsenstein J: PHYLIP - Phylogeny Inference Package (Version 3.2). Cladistics 1989, 5: 164–166.Google Scholar
- Schmid K, Yang Z: The trouble with sliding windows and the selective pressure in BRCA1. PLoS ONE 2008, 3(11):e3746. 10.1371/journal.pone.0003746View ArticlePubMedPubMed CentralGoogle Scholar
- Massingham T, Goldman N: Detecting amino acid sites under positive selection and purifying selection. Genetics 2005, 169(3):1753–1762. 10.1534/genetics.104.032144View ArticlePubMedPubMed CentralGoogle Scholar
- Sawyer SL, Wu LI, Emerman M, Malik HS: Positive selection of primate TRIM5alpha identifies a critical species-specific retroviral restriction domain. Proceedings of the National Academy of Sciences of the United States of America 2005, 102(8):2832–2837. 10.1073/pnas.0409853102View ArticlePubMedPubMed CentralGoogle Scholar
- Zanotto PM, Kallas EG, de Souza RF, Holmes EC: Genealogical evidence for positive selection in the nef gene of HIV-1. Genetics 1999, 153(3):1077–1089.PubMedPubMed CentralGoogle Scholar
- Yang Z: Likelihood ratio tests for detecting positive selection and application to primate lysozyme evolution. Molecular biology and evolution 1998, 15(5):568–573.View ArticlePubMedGoogle Scholar
- Pan C, Kim J, Chen L, Wang Q, Lee C: The HIV positive selection mutation database. Nucleic acids research 2007, (35 Database):D371–375. 10.1093/nar/gkl855Google Scholar
- Stein LD, Bao Z, Blasiar D, Blumenthal T, Brent MR, Chen N, Chinwalla A, Clarke L, Clee C, Coghlan A, et al.: The genome sequence of Caenorhabditis briggsae: a platform for comparative genomics. PLoS biology 2003, 1(2):E45. 10.1371/journal.pbio.0000045View ArticlePubMedPubMed CentralGoogle Scholar
- Francis R, Maine E, Schedl T: Analysis of the multiple roles of gld-1 in germline development: interactions with the sex determination cascade and the glp-1 signaling pathway. Genetics 1995, 139(2):607–630.PubMedPubMed CentralGoogle Scholar
- Haag ES, Ackerman AD: Intraspecific variation in fem-3 and tra-2, two rapidly coevolving nematode sex-determining genes. Gene 2005, 349: 35–42. 10.1016/j.gene.2004.12.051View ArticlePubMedGoogle Scholar
- Clifford R, Lee MH, Nayak S, Ohmachi M, Giorgini F, Schedl T: FOG-2, a novel F-box containing protein, associates with the GLD-1 RNA binding protein and directs male sex determination in the C. elegans hermaphrodite germline. Development (Cambridge, England) 2000, 127(24):5265–5276.Google Scholar
- Lee MH, Schedl T: Identification of in vivo mRNA targets of GLD-1, a maxi-KH motif containing protein required for C. elegans germ cell development. Genes & development 2001, 15(18):2408–2420. 10.1101/gad.915901View ArticleGoogle Scholar
- Nayak S, Goree J, Schedl T: fog-2 and the evolution of self-fertile hermaphroditism in Caenorhabditis. PLoS biology 2005, 3(1):e6. 10.1371/journal.pbio.0030006View ArticlePubMedPubMed CentralGoogle 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.