A novel web-based TinT application and the chronology of the Primate Alu retroposon activity
- Gennady Churakov†1Email author,
- Norbert Grundmann†2,
- Andrej Kuritzin3,
- Jürgen Brosius1,
- Wojciech Makałowski2 and
- Jürgen Schmitz1
© Churakov et al; licensee BioMed Central Ltd. 2010
Received: 18 June 2010
Accepted: 2 December 2010
Published: 2 December 2010
DNA sequences afford access to the evolutionary pathways of life. Particularly mobile elements that constantly co-evolve in genomes encrypt recent and ancient information of their host's history. In mammals there is an extraordinarily abundant activity of mobile elements that occurs in a dynamic succession of active families, subfamilies, types, and subtypes of retroposed elements. The high frequency of retroposons in mammals implies that, by chance, such elements also insert into each other. While inactive elements are no longer able to retropose, active elements retropose by chance into other active and inactive elements. Thousands of such directional, element-in-element insertions are found in present-day genomes. To help analyze these events, we developed a computational algorithm (Transpositions in Transpositions, or TinT) that examines the different frequencies of nested transpositions and reconstructs the chronological order of retroposon activities.
By examining the different frequencies of such nested transpositions, the TinT application reconstructs the chronological order of retroposon activities. We use such activity patterns as a comparative tool to (1) delineate the historical rise and fall of retroposons and their relations to each other, (2) understand the retroposon-induced complexity of recent genomes, and (3) find selective informative homoplasy-free markers of phylogeny. The efficiency of the new application is demonstrated by applying it to dimeric Alu Short INterspersed Elements (SINE) to derive a complete chronology of such elements in primates.
The user-friendly, web-based TinT interface presented here affords an easy, automated screening for nested transpositions from genome assemblies or trace data, assembles them in a frequency-matrix, and schematically displays their chronological activity history.
Discernible transposed elements (TEs) occupy about half of the human genome . They integrate into host DNA in waves of activity. In the face of increasing density, they frequently insert into each other. Nested insertions encrypt valuable historical information about the relative age of the elements, comparable to fossils in distinct layers of earth. As old fossils are absent in young layers, older inactive TEs are not inserted into younger elements. In contrast, young TEs are able to occupy all strata of older elements as well as those active at the same time. Hence TEs active at different historical periods display characteristic insertion profiles. Comprised as they are of a substantial fraction of TEs, mammalian genomes are ideally suited for such analyses. Moreover, even low genomic accumulations (e.g., about 3% genomic coverage of CR1 elements in chicken; ), are sufficient for distinct profiles of retroposon activity .
Over more than one hundred and sixty million years, mammals have accumulated elements from four major classes of transposons, Long INterspersed Elements (LINEs), Short INterspersed Elements (SINEs), retrovirus-like Long Terminal Repeats (LTRs), and DNA transposons . While members of the last group move via a cut and paste mechanism, the other three elements transpose by a copy and paste mechanism via an RNA intermediate reverse transcribed into cDNA. In humans such RNA transposons represent more than 90% of all transposed elements . Active LINE and LTR elements encode the enzymatic machinery that is necessary for their own propagation, and in the case of LINEs also the co-propagation of SINEs or any other RNA. For LINE1-mediated retroposition, there is a slight preference for A-rich integration sites known as kinkable sites . Such regions contain a TTAAAA consensus motif and are frequently found in the junction of dimeric retroposons such as Alu elements in primates. Alu elements are primate-specific, 7SL RNA-derived SINEs that arose from Fossil Left and Right Alu monomers .
A retrospective, sequence-based insight into deep evolutionary periods is feasible via inferences from sequence divergence, but is accompanied by uncertainties due to changing regional and temporal substitution rates, mutation saturation, and the occurrence of highly mutated CpG sites. Especially older, highly diverged, and short elements lead to unreliable estimations. Counting and comparing nested insertions, however, is less sensitive to such considerations.
There are currently two different approaches for calculating the relative activity periods of subtypes of transposed elements, both of which draw on RepeatMasker annotations. The Transposon Cluster Finder (TCF) estimates how often certain elements have been fragmented by the insertions of other elements over evolutionary time . A compilation of representative subsets of interacting transposed elements is then presented in an adjacency matrix displaying frequencies of interruptions optimized for their potential chronological order. This Interruptive Matrix Analysis (IMA) starts from a random chronological order of elements and systematically repositions them so as to minimize the number of nonzero entries in the part of the matrix defined by the artificial transposition of old elements into new ones.
At about the same time as the TCF application was developed, we developed the Transposition in Transposition (TinT) algorithm , which also uses RepeatMasker coordinates to compile interrupted and nested retroposons. The frequencies of fragmented versus nested elements are counted, assembled in a data matrix, and sorted by pre-selected retroposon types. This matrix applies a specific probabilistic likelihood model (Additional file 1) to calculate the relative integration period for each retroposon subtype in relation to all other subtypes.
Due to the high frequency and multiple interactions of different elements, both the TCF and the TinT methods exhibit high intrinsic complexities and are neither easy nor self-explanatorily applicable for the scientific community. To compensate for these shortcomings, we have now developed an easy to use, web-based interface for the TinT application. TinTs can be directly screened for in model organisms or in any allocated RepeatMasker report data. To demonstrate and test the web-based TinT method, we investigated the representative primate genomes of Homo sapiens, Macaca mulatta (rhesus), Callithrix jacchus (New World marmoset), Tarsius syrichta (Tarsius), and Microcebus murinus (grey mouse lemur) and their well-characterized, primate-specific Alu dimeric elements. Because of the well-known evolutionary histories of both the species and their retroposons [8, 9], primates represent an ideal test group for the TinT application.
The TinT application is implemented in a Java environment (version 1.5 or higher) and executed from a bioinformatics web page that runs as an applet on the client computer. TinT reads and optimizes RepeatMasker information of nested transposons and transfers this information into a data matrix of transpositions in transpositions (TinT). The data matrix is than included in a probability calculation to derive a graphical framework of relative activity periods of transposed elements. The probabilistic model considers a simplified assumption with just one period of activity of elements and no specific target site preference. The applet calculates the relative activity periods of elements, but in the current version no time calibration is implemented. The usage of the web-based application is illustrated in Additional file 2.
Principle of the TinT
Elements of type i inserted at time points (k = 1, 2, ..., n i ; where n i is the number of all elements of type i).
In each of such points in time, inserted element of type i may fragment some elements of type j with a certain probability (including the case: j = i).
Considering an identical probability of insertion into any preexisting element, denoted by α, then probability can be represented as , where η j (t) is the number of elements of type j preexisting at time point t.
Function η j (t) is approximated using the normal (Gaussian) distribution with mean t j and standard deviation n j , e.g., , where σ j = n j , and its derivative has the maximum at time point t = t j . All details of the model are presented as Additional file 1.
Web-based version of TinT
The web-based version of TinT is located at http://www.bioinformatics.uni-muenster.de/tools/tint and requires Java version 1.5. The application is written as a java applet and was developed using the multi-language software development environment Eclipse, which is an integrated development environment with a repository system (CVS) in the background that keeps software changes disposable. There are two input options. First, any RepeatMasker report file can be uploaded and variable subsets and combinations of elements can be selected for TinT calculations. Furthermore, pre-analyzed model organisms and specific elements can be selected for a TinT analysis. Currently, 19 pre-computed genomes are available for the TinT analysis and the data (RepeatMasker output files) can be downloaded from http://www.bioinformatics.uni-muenster.de/tools/tint/download/RepeatMasker/.DIR. The TinT activity pattern is then graphically displayed. It should be mentioned that the application is executed locally on the computer where it is accessed. Optional parameters for reading RepeatMasker data may be entered into a special dialog box. Transpositions can be grouped and this information can be loaded from a file to provide flexibility for further analysis. Printing or exporting the generated graphs is a basic part of the software, so the results can be used in other applications. Exporting depends on the standard printer dialog of the computer system - if available the print is directed to a postscript file.
An example data set from primates
From the rhesus monkey (Macaca mulatta) genome sequences we detected 950,960 dimeric elements including 2,008 nested, TinTs. The pattern of these Alu dimers (Figure 2B) is similar to that of human; but includes, in addition, rhesus monkey-specific, AluY-related AluR elements . Because the resolution of individual AluYR elements was too low, the related AluYRb-d elements (TinT-option: merging elements) were combined.
The 2,132 nested elements from the 1,057,994 Alu elements detected in the New World marmoset (Figure 2C) also show a distribution comparable to those of human and rhesus monkey. In addition, there are three New World monkey-specific AluTa elements with the most recent activities .
The 995 nested elements from the 917,008 dimeric Alu elements detected in the Tarsius genomic sequences (Figure 2D) comprise only the two known AluJ elements. The TinT analysis showed that AluJo was older than the AluJb. Tarsius-specific Alu elements have not yet been detected.
The RepeatMasker screen of the gray mouse lemur (Microcebus murinus) sequences (Figure 2E) revealed a total of 438,443 Alu elements, of which 1,905 were unambiguously nested. The mouse lemur lacks the AluJb elements that are present in all other primate groups. In addition to other potential but as yet uncharacterized Alu dimers, there are several recently described, lemur-specific Alu elements with very dominant and recent distributions (AluL, AluLa, and AluMim; Repbase; ).
Comparison to the TCF defragmentation pattern
Giordano et al.  presented their fragmentation analysis based on a Transposon Cluster Finder (TCF) software package. The transposon defragmentation analysis included most known mammalian TE classes and families but only the three main Alu groups AluJ, AluS, and AluY. Therefore a direct comparison to our TinT activity pattern of Alu elements is limited. Furthermore, the TCF software is not freely available to derive a comparable set of data. The TCF pattern for Alu elements roughly confirms the TinT-derived succession of these elements, but indicates an artificial activity overlay of AluJ with AluS and AluY elements.
In light of the many ongoing genome sequence projects, the TinT method should prove to be quite valuable for characterizing the retroposon-influenced architecture and evolutionary history of genomes and provides a basic aid in conducting efficient retroposon-based phylogenetic reconstructions. To test and demonstrate the advanced efficiency of the TinT algorithm and to present a user-friendly web-based application, we performed a comparative analysis of nested primate specific dimeric Alu SINEs, a group of elements with an established evolutionary history [8, 9]. Using standard consensus sequences of Alu repeats  to screen all available genomic sources of primates, represented by human and macaque (both Catarrhini), marmoset (Platyrrhini), Philippine Tarsius (Tarsiiformes), and gray mouse lemur (Strepsirrhini), we extracted and analyzed more than 9,300 nested from 4.5 million detected Alu SINEs. The relative activity periods of Alu elements revealed by the TinT analyses coincide with our current knowledge of these elements in primates .
It should be mentioned that a substantial proportion of the nested elements are ancestral insertion events and consequently are shared among different primate groups. Such common TinTs lead to similar activity patterns of species, especially for older elements (see also for example , Figure 1).
However, comparing the TinT analysis of dimeric Alu elements to the TCF defragmentation pattern  demonstrates that the TinT analysis provides a more accurate activity pattern and implements information about the number of elements in the genome. The TCF defragmentation analysis shows an AluJ activity time span that overlaps with that of both AluS and AluY elements. The TinT profiles clearly indicate that AluJ elements were already silent before the divergence of Anthropoidea and before the appearance of most AluS subtypes and AluY (Additional file 3). Furthermore, the TCF analysis requires that any given element type interact with at least 29% of all other analyzed element types. The TinT model requires that a given element type interact with only two additional types. Especially for small amounts of genomic data, this raises the sensitivity drastically.
For TinT analyses it is important to carefully preselect and compile the elements of interest. Large elements (e.g., the 6,000 bp, full length, primate-specific L1P type of LINEs) have a higher chance of being occupied by other elements than do smaller ones (e.g., the 300 bp Alu SINE elements.) Therefore, we advise users to analyze such groups separately (see for example ). TinT analyses of both monomeric and dimeric elements together should be interpreted with care. Alu elements for example are composed of two monomers connected via an oligo(A) sequence. Such A-rich regions are preferred targets for insertions [5, 19] and can bias the insertion profile.
The insertion patterns of retroposed elements provide a homoplasy-free character set for tracing the evolutionary history of species . The insertion of a given element at the same genomic location in two species and its absence in a reference species indicates a close relationship between the two sharing species . However, randomly choosing retroposed elements for laborious phylogenetic analysis is highly inefficient, while preselecting specific informative element types (e.g., for deep phylogenetic splits) significantly raises the efficiency of downstream experimental analyses. The TinT application provides a priori information about the relative activity periods of given elements (e.g., to investigate old splits by selectively analyzing old elements that were active in the potential common ancestor of a specific group). The phylogenetic application of TinT-derived element activities significantly aided in resolving the evolutionary histories of galliformes , marsupials , and lagomorphs .
It is well known that retroposed elements significantly influence genome evolution, architecture, and gene function; hence, a clear understanding of their insertion events is a key to understanding the genomic architecture of present-day genomes. Therefore, in addition to a statistical compilation of such elements, TinT affords an invaluable tool for analyzing the chronological activity of retroposed elements. Because SINE elements depend on the LINE retroposition machinery for their insertion, their activity is closely connected to LINE activity. As an example, mammalian-wide interspersed elements (MIRs ) coincide with the activity of LINE2 elements and presumably the propagation of MIRs ended after the inactivation of such elements. To understand the dependence of SINE-LINE associations, overlapping activity periods are a first indication of potential interactivity. A potential non-autonomous and autonomous element affiliation was shown for a novel SINE-like snoRTE element and BovB_Plat autonomous retroposons in platypus .
Genome-wide chronological analyses of transposed elements using TinT build on the RepeatMasker detection of elements or fragments thereof. The detection is based on sequence similarity to a predefined compilation of transposons. Although TinT performs a subsequent stringent quality-check of detected fragmented elements, miss-annotations, especially if old and thus highly diverged elements are involved, cannot be completely excluded. Therefore, more sequence data leads to an increase in precision.
In future TinT updates, we plan to implement two additional levels of complexity. By a genome-wide pre-screening of element-specific insertions, we intend to add empirical retroposon information of type-specific (monomer-dimer, short-long elements) insertion probabilities; thus, freeing the algorithm from the assumption that all elements have similar insertion probabilities. Associated with this, we intend to improve the accuracy of TinT analyses by introducing an asymmetric model of element activity, whereby elements will not necessary reach their highest probability of activity at the center of their activity range. Furthermore, we plan to incorporate an absolute time scale of activity by incorporating divergence data of elements.
Required RepeatMasker Input Data
The RepeatMasker source file can be a critical source of errors due to miss-annotations of elements or their fragments. To overcome this potential problem, we developed a quality check of the RepeatMasker TinT coordinates and automatically selected only unambiguously nested insertions for our analyses. The stringent selection works well for genome data and frequently occurring elements, and provides a reliable TinT pattern. For the analysis of lesser quantities of data or genomes with low copy number elements, we have provided the option of applying less stringent parameters (relaxed conditions; Figure 1).
However, for the human genome, the minimal amount of data that is necessary under stringent conditions to retain a full TinT resolution is about 10% of the genome, for instance about 300,000 traces are sufficient to receive the representative full TinT pattern. This means, the pattern is stable and reproducible after adding additional portions of data. It is noteworthy that this calculation varies from species to species and depends on the frequency of available elements. The precision of the TinT approach increases with the amount and quality of the input data. To derive the most reliable TinT pattern, all available sequences of selected species should be downloaded from genome (ftp://ftp.ncbi.nih.gov/genomes/ or trace databases ftp://ftp.ncbi.nih.gov/pub/TraceDB/). The most time-consuming step of the TinT analysis is the upstream RepeatMasker screening. Depending on the amount of genomic data, the size of the RepeatMasker library used, and the available computational power, this process might run for several days. To reduce this screening time, it is advisable to restrict the RepeatMasker library to specific element groups (e.g., SINEs or LINEs). The local RepeatMasker library can be assembled with specific elements or element groups. Similar retroposon types should always be included in one run to avoid artificial annotation of the masked repeats. The report file can be directly applied to downstream processes.
Primate Test Sets of Data
Genomes of Homo sapiens (hg19), Macaca mulatta (rheMac2), and Callithrix jacchus (calJac1) were downloaded from the UCSC Genome Bioinformatics site; http://hgdownload.cse.ucsc.edu/downloads.html; Tarsius syrichta (Tarsyr1.0) and Microcebus murinus (MicMur1.0) genomes were downloaded from the Broad Institute; http://www.broadinstitute.org/ftp/pub/assemblies/mammals.
Availability and Requirements
Project name: TinT
Project home page: http://www.bioinformatics.uni-muenster.de/tools/tint
Operating system: Platform independent (Requires a Java Virtual Machine (JVM) on the target system)
Programming language: Java
Requirements: Java Runtime Environment
License: GPL for academic users
We thank Marsha Bundman for editing the manuscript. The primate paintings were provided by Jón Baldur Hlíðberg. This work was supported by the Deutsche Forschungsgemeinschaft (SCHM 1469/3-1).
- Lander ES, Linton LM, Birren B, Nusbaum C, Zody MC, Baldwin J, Devon K, Dewar K, Doyle M, FitzHugh W, Funke R, Gage D, Harris K, Heaford A, Howland J, Kann L, Lehoczky J, LeVine R, McEwan P, McKernan K, Meldrim J, Mesirov JP, Miranda C, Morris W, Naylor J, Raymond C, Rosetti M, Santos R, Sheridan A, Sougnez C, et al: Initial sequencing and analysis of the human genome. Nature. 2001, 409 (6822): 860-921. 10.1038/35057062.View ArticlePubMedGoogle Scholar
- Hillier LD, Miller W, Birney E, Warren WC, Hardison RC, Ponting CP, Bork P, Burt DW, Groenen MAM, Delany ME, Dodgson JB, Chinwalla AT, Cliften PF, Clifton SW, Delehaunty KD, Fronick C, Fulton RS, Graves TA, Kremitzki C, Layman D, Magrini V, McPherson JD, Miner TL, Minx P, Nash WE, Nhan MN, Nelson JO, Oddy LG, Pohl CS, Randall-Maher J, et al: Sequence and comparative analysis of the chicken genome provide unique perspectives on vertebrate evolution. Nature. 2004, 432 (7018): 695-716. 10.1038/nature03154.View ArticleGoogle Scholar
- Kriegs JO, Matzke A, Churakov G, Kuritzin A, Mayr G, Brosius J, Schmitz J: Waves of genomic hitchhikers shed light on the evolution of gamebirds (Aves: Galliformes). BMC Evol Biol. 2007, 7: 190-10.1186/1471-2148-7-190.PubMed CentralView ArticlePubMedGoogle Scholar
- Kramerov DA, Vassetzky NS: Short retroposons in eukaryotic genomes. Int Rev Cytol. 2005, 247: 165-221. 10.1016/S0074-7696(05)47004-7.View ArticlePubMedGoogle Scholar
- Jurka J, Klonowski P, Trifonov EN: Mammalian retroposons integrate at kinkable DNA sites. J Biomol Struct Dyn. 1998, 15 (4): 717-721.View ArticlePubMedGoogle Scholar
- Quentin Y: Origin of the Alu family: a family of Alu-like monomers gave birth to the left and the right arms of the Alu elements. Nucleic Acids Res. 1992, 20 (13): 3397-3401. 10.1093/nar/20.13.3397.PubMed CentralView ArticlePubMedGoogle Scholar
- Giordano J, Ge Y, Gelfand Y, Abrusan G, Benson G, Warburton PE: Evolutionary history of mammalian transposons determined by genome-wide defragmentation. PLoS Comput Biol. 2007, 3 (7): e137-10.1371/journal.pcbi.0030137.PubMed CentralView ArticlePubMedGoogle Scholar
- Britten RJ: Rates of DNA sequence evolution differ between taxonomic groups. Science. 1986, 231 (4744): 1393-1398. 10.1126/science.3082006.View ArticlePubMedGoogle Scholar
- Goodman M, Porter CA, Czelusniak J, Page SL, Schneider H, Shoshani J, Gunnell G, Groves CP: Toward a phylogenetic classification of Primates based on DNA evidence complemented by fossil evidence. Mol Phylogenet Evol. 1998, 9 (3): 585-598. 10.1006/mpev.1998.0495.View ArticlePubMedGoogle Scholar
- Han K, Konkel MK, Xing J, Wang H, Lee J, Meyer TJ, Huang CT, Sandifer E, Hebert K, Barnes EW, Hubley R, Miller W, Smit AF, Ullmer B, Batzer MA: Mobile DNA in Old World monkeys: a glimpse through the rhesus macaque genome. Science. 2007, 316 (5822): 238-240. 10.1126/science.1139462.View ArticlePubMedGoogle Scholar
- Ray DA, Batzer MA: Tracking Alu evolution in New World primates. BMC Evol Biol. 2005, 5: 51-10.1186/1471-2148-5-51.PubMed CentralView ArticlePubMedGoogle Scholar
- Liu GE, Alkan C, Jiang L, Zhao S, Eichler EE: Comparative analysis of Alu repeats in primate genomes. Genome Res. 2009, 19 (5): 876-885. 10.1101/gr.083972.108.PubMed CentralView ArticlePubMedGoogle Scholar
- Jurka J, Milosavljevic A: Reconstruction and analysis of human Alu genes. J Mol Evol. 1991, 32 (2): 105-121. 10.1007/BF02515383.View ArticlePubMedGoogle Scholar
- Kapitonov V, Jurka J: The age of Alu subfamilies. J Mol Evol. 1996, 42 (1): 59-65. 10.1007/BF00163212.View ArticlePubMedGoogle Scholar
- Nilsson MA, Churakov G, Sommer M, Tran NV, Zemann A, Brosius J, Schmitz J: Tracking marsupial evolution using archaic genomic retroposon insertions. PLoS Biol. 2010, 8 (7): e1000436-10.1371/journal.pbio.1000436.PubMed CentralView ArticlePubMedGoogle Scholar
- Schmitz J, Ohme M, Zischler H: SINE insertions in cladistic analyses and the phylogenetic affiliations of Tarsius bancanus to other primates. Genetics. 2001, 157 (2): 777-784.PubMed CentralPubMedGoogle Scholar
- Schmitz J, Roos C, Zischler H: Primate phylogeny: molecular evidence from retroposons. Cytogenet Genome Res. 2005, 108 (1-3): 26-37. 10.1159/000080799.View ArticlePubMedGoogle Scholar
- Warren WC, Hillier LW, Marshall Graves JA, Birney E, Ponting CP, Grutzner F, Belov K, Miller W, Clarke L, Chinwalla AT, Yang SP, Heger A, Locke DP, Miethke P, Waters PD, Veyrunes F, Fulton L, Fulton B, Graves T, Wallis J, Puente XS, Lopez-Otin C, Ordonez GR, Eichler EE, Chen L, Cheng Z, Deakin JE, Alsop A, Thompson K, Kirby P, et al: Genome analysis of the platypus reveals unique signatures of evolution. Nature. 2008, 453 (7192): 175-183. 10.1038/nature06936.PubMed CentralView ArticlePubMedGoogle Scholar
- Levy A, Schwartz S, Ast G: Large-scale discovery of insertion hotspots and preferential integration sites of human transposed elements. Nucleic Acids Res. 2009, 38 (5): 1515-1530. 10.1093/nar/gkp1134.PubMed CentralView ArticlePubMedGoogle Scholar
- Ryan SC, Dugaiczyk A: Newly arisen DNA repeats in primate phylogeny. Proc Natl Acad Sci USA. 1989, 86 (23): 9360-9364. 10.1073/pnas.86.23.9360.PubMed CentralView ArticlePubMedGoogle Scholar
- Shedlock AM, Okada N: SINE insertions: powerful tools for molecular systematics. Bioessays. 2000, 22 (2): 148-160. 10.1002/(SICI)1521-1878(200002)22:2<148::AID-BIES6>3.0.CO;2-Z.View ArticlePubMedGoogle Scholar
- Kriegs JO, Zemann A, Churakov G, Matzke A, Ohme M, Zischler H, Brosius J, Kryger U, Schmitz J: Retroposon insertions provide insights into deep lagomorph evolution. Mol Biol Evol. 2010Google Scholar
- Smit AF, Riggs AD: MIRs are classic, tRNA-derived SINEs that amplified before the mammalian radiation. Nucleic Acids Res. 1995, 23 (1): 98-102. 10.1093/nar/23.1.98.PubMed CentralView ArticlePubMedGoogle Scholar
- Schmitz J, Zemann A, Churakov G, Kuhl H, Grutzner F, Reinhardt R, Brosius J: Retroposed SNOfall - A mammalian-wide comparison of platypus snoRNAs. Genome Research. 2008, 18 (6): 1005-1010. 10.1101/gr.7177908.PubMed CentralView ArticlePubMedGoogle Scholar
- Steiper ME, Young NM: Primate molecular divergence dates. Mol Phylogenet Evol. 2006, 41 (2): 384-394. 10.1016/j.ympev.2006.05.021.View ArticlePubMedGoogle Scholar
- Roos C, Schmitz J, Zischler H: Primate jumping genes elucidate strepsirrhine phylogeny. Proc Natl Acad Sci USA. 2004, 101 (29): 10650-10654. 10.1073/pnas.0403852101.PubMed CentralView ArticlePubMedGoogle 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://www.creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.