Vol 437|20 October 2005|doi:10.1038\/nature04107
\nLETTERS
\nAdaptive evolution of non-coding DNA in Drosophila
\nPeter Andolfatto1
\nA large fraction of eukaryotic genomes consists of DNA that is not
\ntranslated into protein sequence, and little is known about its
\nfunctional significance. Here I show that several classes of noncoding DNA in Drosophila are evolving considerably slower than
\nsynonymous sites, and yet show an excess of between-species
\ndivergence relative to polymorphism when compared with
\nsynonymous sites. The former is a hallmark of selective constraint,
\nbut the latter is a signature of adaptive evolution, resembling
\ngeneral patterns of protein evolution in Drosophila1,2. I estimate
\nthat about 40\u201370% of nucleotides in intergenic regions, untranslated portions of mature mRNAs (UTRs) and most intronic DNA
\nare evolutionarily constrained relative to synonymous sites. However, I also use an extension to the McDonald\u2013Kreitman test3 to
\nshow that a substantial fraction of the nucleotide divergence in
\nthese regions was driven to fixation by positive selection (about
\n20% for most intronic and intergenic DNA, and 60% for UTRs).
\nOn the basis of these observations, I suggest that a large fraction of
\nthe non-translated genome is functionally important and subject
\nto both purifying selection and adaptive evolution. These results
\nimply that, although positive selection is clearly an important facet
\nof protein evolution, adaptive changes to non-coding DNA might
\nhave been considerably more common in the evolution of
\nD. melanogaster.
\nThe high degree of protein sequence similarity between phenotypically diverged species has led some to propose that regulatory
\nevolution may be of considerably more importance than protein
\nevolution4,5. Although most of the typical eukaryotic genome is
\ncomprised of non-coding DNA, comparatively little is known
\nabout the evolutionary forces acting on it. Some unknown fraction
\nof the non-translated genome is presumed to be crucial for the
\nregulation of gene expression. Most of our direct knowledge regarding the evolution of regulatory elements comes from a handful of
\ndirect functional studies5,6. A second, indirect approach is based on
\ncomparative genomics7
\n. The rationale for this second approach is
\nthat if newly arising mutations are typically detrimental to gene
\nfunction, functionally important parts of the genome are expected to
\nevolve more slowly than those lacking function8\u201311.
\nThere are some limitations to the comparative genomics
\napproach. First, a given genomic region might be conserved owing
\nsimply to a lower mutation rate12. Second, known regulatory
\nelements do not seem to be particularly well conserved as a class,
\nat least in Drosophila10. This fnding suggests that taking an approach
\nbased on sequence conservation alone may lead to a biased view of
\nregulatory evolution. Functionality of DNA sequences implies that
\nthey can be subject to both negative and positive selection. If a
\nsignifcant fraction of divergence between species observed in noncoding DNA is positively selected rather than selectively neutral or
\nconstrained, this could lead to underestimates of the functional
\nimportance of non-coding DNA and cause researchers to overlook
\nthe contribution of arguably the most interesting class of mutations
\nin genome evolution\u2014those refecting adaptive differences between
\npopulations and species.
\nThese limitations can be overcome by combining comparative
\ngenomic analyses with population-level variability data1\u20133,13. To
\nassess the mode of selection acting on non-coding DNA, I have
\nanalysed new and previously published polymorphism data for 35
\ncoding fragments (average length 667 base pairs (bp)) and 153 noncoding fragments (average length 426 bp) scattered across the
\nX chromosome of D. melanogaster (see Supplementary Materials 1).
\nTo estimate levels of between-species divergence, I have compared
\nD. melanogaster with its closely related sibling species, D. simulans.
\nOn the basis of the current Drosophila genome annotation (release
\n4), I separated the surveyed fragments into several categories that are
\nlikely to differ in the intensity and mode of selection acting on them
\n(see Table 1). It is apparent that most non-coding DNA evolves
\nconsiderably slower than synonymous sites (that is, sites in proteincoding sequences at which mutations do not result in amino acid
\nsubstitutions; Table 1). This is the case for introns and UTRs (see also
\nrefs 14\u201316), as well as intergenic DNA, much of which is far from the
\nclosest known gene (see Supplementary Materials 1). I estimate levels
\nof constraint in Drosophila non-coding DNA to be 40% for introns,
\n50% for intergenic regions (IGRs), and 60% for UTRs (Table 2).
\nThese are all considerably higher than previous estimates from a
\nvariety of species comparisons11,15\u201318 . The non-coding DNA
\nsurveyed is also generally less polymorphic than synonymous sites
\nin D. melanogaster (Table 1; p , 10210, Wilcoxon two-sample test for
\nUTRs and introns\u00feIGRs versus synonymous sites). Thus, both
\npolymorphism and divergence in non-coding DNA are signifcantly
\nreduced relative to synonymous sites in D. melanogaster.
\nReduced levels of polymorphism and divergence in non-coding
\nDNA resemble general patterns of protein evolution19 and suggest
\nthat non-coding DNA is either functionally constrained or is subject
\nto a lower mutation rate than synonymous sites. One way to
\ndistinguish between these two models is to consider the distribution
\nof polymorphism frequencies. Negative selection acting on polymorphic variants will keep them at lower frequencies in a population
\nthan expected if they were neutral20. Consistent with this prediction,
\nthe distribution of polymorphism frequencies at both non-coding
\nDNA and amino acid sites is skewed towards rare frequencies relative
\nto synonymous polymorphisms (as indicated by a more negative
\nTajima\u2019s D value20, Fig. 1). The distribution of Tajima\u2019s D values for
\nnon-synonymous sites among loci is negatively skewed relative to
\nsynonymous sites, suggesting that amino acid polymorphisms are
\nsubject to purifying selection (Fig. 1; p \u00bc 0.002, Wilcoxon twosample test versus synonymous sites). Here I show that this
\nsame pattern extends to polymorphisms in non-coding DNA
\n(Fig. 1; Wilcoxon test versus synonymous sites: pooled non-coding,
\np \u00bc 0.0001; UTRs, p , 0.0001; introns, p \u00bc 0.001; IGRs, p \u00bc 0.005).
\nThis fnding, together with the observed reduction in polymorphism
\nand divergence, implies that mutations in non-coding DNA are
\nsubject, on average, to stronger negative selection than synonymous
\nsites (see also Supplementary Materials 2).
\nDoes selective constraint alone account for patterns of non-coding
\nDNA evolution? McDonald and Kreitman3 have proposed a frame1
\nSection of Ecology, Behavior and Evolution, Division of Biological Sciences, University of California San Diego, La Jolla, California 92093, USA.
\n1149
\n\u00a9 2005 Nature Publishing Group
\nLETTERS NATURE|Vol 437|20 October 2005
\nTable 1 | Polymorphism and divergence in coding and non-coding DNA of D. melanogaster
\nMutation class No. of regions Mean p* Mean Dxy\u2020 D\u2021 P\u00a7 pk P\u2019{ p#
\nSynonymous 35 2.87 13.59 604 502 2 323 2
\nNon-synonymous 35 0.18 1.72 260 115 ,1026 52 ,1029
\nNon-coding 153 1.06 5.94 3,168 2,386 0.14 1,295 ,1023
\nUTRs 31 0.54 4.54 471 246 ,1025 107 ,10211
\n50
\nUTRs 18 0.61 5.41 328 160 ,1025 71 ,1029
\n30
\nUTRs 13 0.45 3.35 143 86 0.034 36 ,1024
\nIntrons 72 1.25 6.71 1,564 1,221 0.39 675 0.010
\nIGRs 50 1.11 5.72 1,133 919 .0.5 513 0.059
\npIGRs 20 1.29 6.58 500 400 .0.5 237 0.25
\ndIGRs 30 0.99 5.18 633 519 .0.5 276 0.041
\nIntrons\u00feIGR 122 1.19 6.25 2,697 2,140 0.50 1,188 0.013
\nMutation classes: synonymous sites, non-synonymous sites, untranslated transcribed regions (UTRs), intergenic regions within 2 kb of a gene (pIGRs), intergenic regions more than 4 kb away
\nfrom a gene (dIGRs).
\n*p is the weighted average within-species pairwise diversity per 100 sites.
\n\u2020Dxy is the weighted average pairwise divergence per 100 sites between D. melanogaster and D. simulans, corrected for multiple hits (Jukes\u2013Cantor). Dxy at fourfold degenerate synonymous
\nsites is 12.0%.
\n\u2021D is the estimated number of fxed differences between species using a Jukes\u2013Cantor correction for multiple hits (see Methods).
\n\u00a7P is the number of intraspecifc polymorphisms.
\nkMcDonald\u2013Kreitman test of probability using all polymorphisms.
\n{P\u2019 is the number of intraspecifc polymorphisms excluding singletons.
\n#McDonald\u2013Kreitman test of probability excluding singleton polymorphisms. Probabilities are from two-tailed Fisher\u2019s exact tests and assume sites are independent. These are likely to be only
\nslight underestimates given probable levels of intragenic recombination (see Supplementary Materials 2).
\nwork to distinguish neutrality (and variation in mutation rate) from
\nnegative and positive selection in the genome. Their approach
\ncompares levels of polymorphism within and divergence between
\nspecies for a putatively selected class of sites in the genome to a
\nneutral standard. If reduced levels of polymorphism and divergence
\nin non-coding DNA can be explained by a lower mutation rate, the
\nratio of polymorphism to divergence should be similar to that for
\nsynonymous sites. Positive selection will increase divergence relative
\nto polymorphism at selected sites, whereas negative selection is
\nexpected to result in the opposite pattern21. Although this framework
\nwas originally designed to detect selection within protein-coding
\ngenes, it can be generalized to consider arbitrary classes of putatively
\nselected sites sampled from multiple genomic regions, including
\nnon-coding DNA (see Supplementary Materials 2). Using all polymorphisms, there is a signifcant excess of divergence for amino
\nacid replacement sites (p \u00bc 5 \u00a3 1027
\n) and for UTRs (p \u00bc 3 \u00a3 1026
\n,
\ntwo-tailed Fisher\u2019s exact test) but not at other subclasses of noncoding DNA (Table 1). This preliminary analysis suggests that,
\nsimilar to the pattern observed for amino acid substitutions1,2,a
\nsignifcant proportion of nucleotide divergence at UTRs was also
\ndriven to fxation by positive selection.
\nThe presence of weakly negatively selected variants in polymorphism can mask the signature of adaptive evolution in the
\ngenome1,22, making the McDonald\u2013Kreitman test very conservative.
\nAs I have shown above that polymorphic variants in non-coding
\nDNA are subject to stronger selective constraint than synonymous
\nsites (Table 1 and Fig. 1), negatively selected variants contributing to
\npolymorphism in non-coding DNA are likely to be a factor limiting
\nTable 2 | Functionally relevant nucleotides in non-coding DNA
\nClass C (%)* a(%)\u2020 p (a # 0)\u2021 FRN (%)\u00a7
\nUTRs 60.4 57.5 ,1023 83.2
\n50
\nUTRs 52.9 60.8 ,1023 80.9
\n30
\nUTRs 70.7 52.9 ,1023 86.2
\nIntrons 39.5 19.3 0.007 51.2
\nIGRs 49.3 15.3 0.036 57.1
\npIGRs 40.6 11.4 0.165 47.4
\ndIGRs 54.6 18.5 0.019 63.0
\nIntrons \u00fe IGR 44.2 17.6 0.013 54.0
\n*Constraint (C) is estimated relative to fourfold degenerate synonymous sites.
\n\u2020a is the estimated fraction of divergence driven by positive selection.
\n\u2021Probabilities (a # 0) have been adjusted for effects of linkage within loci (see
\nSupplementary Materials 2.5).
\n\u00a7FRN is the inferred fraction of functionally relevant nucleotides given levels of constraint
\nand a (that is, FRN < C \u00fe (1 2 C)a).
\n1150
\npower to detect positive selection. This problem can be partially
\novercome by considering only those mutations that are not rare
\nin a sample from both the neutral and putatively selected classes
\n(see ref. 23 and Supplementary Materials 2). Applying this approach
\nreveals a signifcant excess of divergence in UTRs and in most
\nother classes of non-coding DNA relative to synonymous sites
\n(Table 1; UTRs, p \u00bc 5 \u00a3 10212; introns, p \u00bc 0.01; dIGRs, p \u00bc 0.04;
\nintrons\u00feIGRs, p \u00bc 0.01). A Hudson\u2013Kreitman\u2013Aguade\u00b4 (HKA)
\ntest24 also provides statistical support for a reduced ratio of polymorphism to divergence for non-coding DNA relative to synonymous sites (UTRs, p , 1023
\n; pooled introns and IGRs, p \u00bc 0.02; see
\nSupplementary Materials 2). Together, these results show that a
\nsignifcant fraction of the divergence in UTRs, introns and intergenic
\nDNA was probably driven to fxation by positive selection.
\nTo quantify the intensity and the relative importance of positive
\nselection in shaping the evolution of non-coding DNA, I apply two
\nextensions of the McDonald\u2013Kreitman approach2,13. First I estimate
\na, defned as the proportion of the divergence between species that
\nwas driven by positive selection2
\n. I estimate that about 20% of the
\nnucleotide divergence in introns and intergenic DNA was driven to
\nfxation by positive selection, and about 60% for UTRs (Fig. 2a and
\nTable 2). Using a hierarchical bayesian framework13, I estimate the<\/p>\n
Figure 1 | Mean Tajima\u2019s D values for coding and non-coding DNA. Means
\nacross loci are given with bars indicating two standard errors. The
\nexpectation of D under the neutral model is shown as a dotted line. Syn,
\nsynonymous sites; NonSyn, non-synonymous sites; NonCod, pooled
\nnon-coding DNA.
\n\u00a9 2005 Nature Publishing Group
\nNATURE|Vol 437|20 October 2005 LETTERS
\nselection intensity on non-coding DNA (including UTRs, introns
\nand IGRs) to be positive and signifcantly different from zero in most
\ncases (Fig. 2b; Supplementary Materials 3). As this bayesian approach
\nassumes that segregating and fxed variants are subject to the
\nsame direction and intensity of selection, it is likely to underestimate
\nthe magnitude of 2Nes (the intensity of selection) for nucleotide
\nsubstitutions fxed by positive selection (see Supplementary
\nMaterials 2).
\nEvidence that a signifcant fraction of non-coding DNA is functionally important is emerging from a variety of comparative
\ngenomic studies. However, my fnding of a large fraction of positively
\nselected divergence implies that \u2018evolutionary constraint\u2019 will substantially underestimate the fraction of functionally relevant nucleotides because it ignores the contribution of positively selected
\nmutations to divergence. For the example of UTRs, I estimate
\nevolutionary constraint to be 60%. However, as 58% of the observed
\ndivergence was positively selected, this implies that 83% of nucleotides in UTRs are in fact functionally relevant. Likewise, the fraction
\nof functionally relevant nucleotides in introns and IGRs is likely to be
\nabout 10\u201320% higher than suggested by levels of constraint alone
\n(Table 2).
\nHow frequent is adaptation in the Drosophila genome? Rough
\ncalculations (see Supplementary Materials 4) suggest that there has
\nbeen about one adaptive amino acid substitution every 20 years since
\nthe split of D. melanogaster and D. simulans (see also ref. 2). Although
\nthis is substantial, consider that the total number of sites contained in<\/p>\n
Figure 2 | Quantifying adaptive divergence and selection intensity.
\na, Estimates of a, the fraction of nucleotide divergence driven by positive
\nselection. Error bars indicate 90% confidence limits determined by a nonparametric bootstrapping. Estimated probabilities that a $ 0 corrected for
\npartial linkage are given in Table 2. b, Estimates of the intensity of selection
\n(2Nes) acting on non-synonymous and non-coding DNA sites. Error bars
\nindicate 90% confidence limits determined by simulation (see Methods).
\nSingleton polymorphisms were excluded in estimates of a and 2Nes (see
\nSupplementary Materials 3). Abbreviations as in Fig. 1.
\nintrons, intergenic regions and UTRs far outweighs the number of
\ncodons in the Drosophila genome25. I estimate that UTRs alone
\ncontribute as much to adaptive divergence between species as do
\namino acid changes, and the summed contribution of non-coding
\nDNA to adaptive divergence could easily be an order of magnitude
\nlarger. These fndings support previous intuitions4,5 about the great
\nimportance of regulatory changes in evolution.
\nMETHODS
\nData. All loci used in this study, previously published or newly collected, are
\nX-linked genomic fragments, with a sample size of 12 D. melanogaster alleles
\nsampled from a population in Zimbabwe, and a single D. simulans sequence.
\nFor coding DNA (synonymous and non-synonymous sites), I collected polymorphism and divergence in 31 coding regions selected randomly with respect
\nto gene function, and 51 non-coding regions (27 intergenic and 24 untranslated
\ntranscribed regions). Information about these 82 loci and primers used can be
\nfound in Supplementary Materials 1. I used polymerase chain reaction (PCR) to
\namplify 700\u2013800-bp regions from genomic DNA extracted from single male fies,
\nremoved primers and nucleotides using exonuclease I and shrimp alkaline
\nphosphatase, and sequenced the cleaned product on both strands using Big-Dye
\n(Version 3, Applied Biosystems). Sequences were collected on an ABI 3730
\ncapillary sequencer and were aligned and edited using the program Sequencher
\n(Gene Codes).
\nTo the 82 regions surveyed above, I added previously published data for
\nloci that had the same sample size (n \u00bc 12 fies) and were surveyed in similar
\nsamples from Zimbabwe26,27. A number of the previously published loci26 had
\nto be functionally reassigned when compared to Release 4 of the annotated
\nD. melanogaster genome (http:\/\/fybase.bio.indiana.edu\/annot\/dmel-release4.
\nhtml). I excluded any loci in regions of reduced recombination (see below).
\nPreviously published loci ftting these requirements were processed into 106
\nfragments (4 coding, 7 UTR, 23 intergenic and 72 intron). Thus, the total
\nnumber of regions surveyed in this analysis is 188. Alignments for each locus
\nare available upon request. A reciprocal best-hit BLAST protocol was used to
\nconfrm that the regions compared between D. melanogaster and D. simulans
\nare indeed orthologous. Extra gaps were introduced into some alignments in
\nregions that were particularly diffcult to align. This procedure is likely to
\nupwardly bias estimates of constraint, but is conservative with respect to
\ndetecting positive selection.
\nAnalyses. The estimated number of synonymous sites, non-synonymous sites,
\naverage pairwise diversity (p), average pairwise divergence (Dxy), as well as
\ncounts of the number of polymorphic sites (P) were performed using DnaSP
\nsoftware (version 4; http:\/\/www.ub.es\/dnasp\/) and Perl code written by P.A. The
\nnumber of divergent sites (D) was estimated as Dxy 2 p using a Jukes\u2013Cantor
\ncorrection for multiple hits. Multiply hit sites were included in the analysis but
\ninsertion\u2013deletion polymorphisms and mutations overlapping alignment gaps
\nwere excluded. Derived mutations were polarized using a single D. simulans
\nsequence and assuming standard parsimony criteria. Tajima\u2019s D value20 was
\nestimated from the number of polymorphisms and p.
\nIn this study, I assume that synonymous sites are more neutral than putatively
\nselected classes of sites (see Supplementary Materials 2.2). I separated noncoding DNA into subclasses that I expected a priori to experience different
\nselection pressures: 50 and 30
\nuntranslated transcribed regions (UTRs), introns,
\nintergenic regions within 2 kilobases (kb) of a gene (proximal intergenic regions,
\npIGRs), and intergenic regions further than 4 kb from the nearest gene (distal
\nintergenic regions, dIGRs). My sample of intron fragments is biased towards
\nintrons larger than the median intron size (86 bp) (ref. 28), making estimates of
\nconstraint higher than expected with a random sample of introns14. However,
\n95% of intronic DNA is contained within introns longer than the median size28,
\nand thus my estimate refects levels of constraint for most intronic DNA in the
\ngenome.
\nFor comparisons of polymorphism and divergence between synonymous
\nsites and non-coding DNA, it was necessary to pool sites in each class. I estimate
\nevolutionary constraint relative to fourfold degenerate synonymous sites using
\nthe approach in ref. 15, except that I pooled classes of sites and used a Jukes\u2013
\nCantor correction for multiple hits19. Given differences in base composition
\nbetween coding and non-coding regions, I investigated possible differences in
\nmutations rates owing to the 16 possible adjacent-base contexts of nucleotides
\n(suggested by A. Kondrashov). There was no signifcant effect of adjacent-base
\ncontext on rates of divergence (see Supplementary Materials 5).
\nI estimate the proportion of divergence driven by positive selection1,2 as
\na \u00bc 1\u2013(DSPX\/DXPS), where S denotes synonymous (that is, putatively neutral)
\nP P n n sites, X denotes putatively selected sites, and D \u00bc i\u00bc1 and P \u00bc Pi D , i i\u00bc1
\n1151
\n\u00a9 2005 Nature Publishing Group<\/p>\n
LETTERS NATURE|Vol 437|20 October 2005
\nwhere Di and Pi are the number of divergent and polymorphic variants at locus i,
\nrespectively, and n is the number of loci of class S or X. Confdence limits on a
\nwere estimated using a standard non-parametric bootstrapping procedure,
\nassuming sites are independent. The issue of non-independence of sites within
\nsurveyed fragments is addressed in Supplementary Materials 2.5. For consistency, a was estimated for non-synonymous sites in the same way. The intensity
\nof selection (2Nes) was estimated on putatively selected classes (pooling sites as
\nabove) using a hierarchical bayesian method (http:\/\/cbsuapps.tc.cornell.edu)13.
\nTo avoid problems associated with large-scale variation in recombination rates, I
\nrestricted my survey of loci to regions of the X chromosome that have the highest
\nrates of recombination29 (see Supplementary Fig. 1.1).
\nReceived 23 May; accepted 2 August 2005.
\n1. Fay, J. C., Wyckoff, G. J. & Wu, C. I. Testing the neutral theory of molecular
\nevolution with genomic data from Drosophila. Nature 415, 1024\u2013-1026
\n(2002).
\n2. Smith, N. G. & Eyre-Walker, A. Adaptive protein evolution in Drosophila. Nature
\n415, 1022\u2013-1024 (2002).
\n3. McDonald, J. & Kreitman, M. Adaptive protein evolution at the Adh locus in
\nDrosophila. Nature 351, 652\u2013-654 (1991).
\n4. King, M. C. & Wilson, A. C. Evolution at two levels in humans and
\nchimpanzees. Science 188, 107\u2013-116 (1975).
\n5. Carroll, S. B., Grenier, J. K. & Weatherbee, S. D. From DNA to Diversity:
\nMolecular Genetics and the Evolution of Animal Design (Blackwell Science,
\nMalden, Massachusetts, 2001).
\n6. Ludwig, M. et al. Functional evolution of a cis-regulatory module. PLoS Biol. 3,
\ne93 (2005).
\n7. Miller, W., Makova, K., Nekrutenko, A. & Hardison, R. Comparative genomics.
\nAnnu. Rev. Genomics Hum. Genet. 5, 15\u2013-56 (2004).
\n8. Cliften, P. et al. Surveying Saccharomyces genomes to identify functional
\nelements by comparative DNA sequence analysis. Genome Res. 11, 1175\u2013-1186
\n(2001).
\n9. Gibbs, R. et al. Genome sequence of the Brown Norway rat yields insights into
\nmammalian evolution. Nature 428, 493\u2013-521 (2004).
\n10. Richards, S. et al. Comparative genome sequencing of Drosophila pseudoobscura:
\nchromosomal, gene, and cis-element evolution. Genome Res. 15, 1\u2013-18 (2005).
\n11. Shabalina, S. & Kondrashov, A. Pattern of selective constraint in C. elegans and
\nC. briggsae genomes. Genet. Res. 74, 23\u2013-30 (1999).
\n12. Clark, A. The search for meaning in noncoding DNA. Genome Res. 11,
\n1319\u2013-1320 (2001).
\n13. Bustamante, C. et al. The cost of inbreeding in Arabidopsis. Nature 416,
\n531\u2013-534 (2002).
\n14. Haddrill, P. R., Halligan, D., Charlesworth, B. & Andolfatto, P. Patterns of intron
\nsequence evolution in Drosophila are dependent upon length and GC content.
\nGenome Biol. 6, R67 (2005).
\n15. Halligan, D., Eyre-Walker, A., Andolfatto, P. & Keightley, P. Patterns of
\nevolutionary constraints in intronic and intergenic DNA of Drosophila. Genome
\nRes. 14, 273\u2013-279 (2004).
\n16. Bachtrog, D. Sex chromosome evolution: molecular aspects of Y chromosome
\ndegeneration in Drosophila. Genome Res. 15, 1393\u2013-1401 (2005).
\n17. Jareborg, N., Birney, E. & Durbin, R. Comparative analysis of noncoding regions
\nof 77 orthologous mouse and human gene pairs. Genome Res. 9, 815\u2013-824
\n(1999).
\n18. Bergman, C. & Kreitman, M. Analysis of conserved noncoding DNA in
\nDrosophila reveals similar constraints in intergenic and intronic sequences.
\nGenome Res. 11, 1335\u2013-1345 (2001).
\n19. Li, W. Molecular Evolution (Sinauer Associates, Sunderland, Massachusetts,
\n1997).
\n20. Tajima, F. Statistical method for testing the neutral mutation hypothesis by
\nDNA polymorphism. Genetics 123, 585\u2013-595 (1989).
\n21. Kimura, M. The Neutral Theory of Molecular Evolution (Cambridge Univ. Press,
\nCambridge, 1983).
\n22. Charlesworth, B. The effect of background selection against deleterious
\nmutations on weakly selected, linked variants. Genet. Res. 63, 213\u2013-227 (1994).
\n23. Templeton, A. Contingency tests of neutrality using intra\/interspecifc gene
\ntrees: the rejection of neutrality for the evolution of the mitochondrial
\ncytochrome oxidase II gene in the hominoid primates. Genetics 144, 1263\u2013-1270
\n(1996).
\n24. Hudson, R., Kreitman, M. & Aguade\u00b4, M. A test of neutral molecular evolution
\nbased on nucleotide data. Genetics 116, 153\u2013-159 (1987).
\n25. Misra, S. et al. Annotation of the Drosophila melanogaster euchromatic genome:
\na systematic review. Genome Biol. 3, research0083.1-0083.22 (2002).
\n26. Glinka, S., Ometto, L., Mousset, S., Stephan, W. & De Lorenzo, D. Demography
\nand natural selection have shaped genetic variation in Drosophila melanogaster:
\na multi-locus approach. Genetics 165, 1269\u2013-1278 (2003).
\n27. Haddrill, P. R., Thornton, K. R., Charlesworth, B. & Andolfatto, P. Multilocus
\npatterns on nucleotide variability and the demographic and selection history of
\nDrosophila melanogaster populations. Genome Res. 15, 790\u2013-799 (2005).
\n28. Yu, J. et al. Minimal introns are not \u201cjunk\u201d. Genome Res. 12, 1185\u2013-1189 (2002).
\n29. Charlesworth, B. Background selection and patterns of genetic diversity in
\nDrosophila melanogaster. Genet. Res. 68, 131\u2013-149 (1996).
\nSupplementary Information is linked to the online version of the paper at
\nwww.nature.com\/nature.
\nAcknowledgements The author thanks D. Bachtrog for extensive comments on
\nthe manuscript and help with data quality issues, C. Bustamante and K. Thornton
\nfor providing code, and B. Ballard for Zimbabwe fy lines. P. Haddrill and
\nK. Thornton Helped in designing primers for distal intergenic and coding
\nregions, respectively. Thanks to B. Fischman for technical help, A. Betancourt,
\nA. Kondrashov, A. Poon, D. Presgraves, M. Przeworski and S. Wright for critical
\ncomments on the manuscript, and L. Chao and J. Huelsenbeck for advice.
\nThanks also to the Washington University Genome Sequencing Center for
\nproviding unpublished D. simulans sequences. This work was funded in part by a
\nresearch grant from the Biotechnology and Biological Sciences Research Council
\n(UK) to P.A. The author is supported by an Alfred P. Sloan Fellowship in
\nMolecular and Computational Biology.
\nAuthor Information Reprints and permissions information is available at
\nnpg.nature.com\/reprintsandpermissions. The author declares no competing
\nfnancial interests. Correspondence and requests for materials should be
\naddressed to P.A. (pandolfatto@ucsd.edu).
\n1152
\n\u00a9 2005 Nature Publishing Group<\/p>\n
—–<\/p>\n
10.1038\/nature04107|Volume 437|20 October 2005|doi:10.1038\/nature04107<\/p>\n
LETTERS<\/p>\n
Drosophila’s non-coding DNA has evolved in a way that is adaptive.<\/p>\n
Andolfatto, Peter1<\/p>\n
A large fraction of eukaryotic genomes consists of DNA that is not<\/p>\n
translated into protein sequence, and little is known about its<\/p>\n
functional significance. Here I show that several classes of noncoding DNA in Drosophila are evolving considerably slower than<\/p>\n
synonymous sites, and yet show an excess of between-species<\/p>\n
divergence relative to polymorphism when compared with<\/p>\n
synonymous sites. The former is a hallmark of selective constraint,<\/p>\n
but the latter is a signature of adaptive evolution, resembling<\/p>\n
general patterns of protein evolution in Drosophila1,2. I estimate<\/p>\n
that about 40\u201370% of nucleotides in intergenic regions, untranslated portions of mature mRNAs (UTRs) and most intronic DNA<\/p>\n
are evolutionarily constrained relative to synonymous sites. However,<\/p>\n","protected":false},"excerpt":{"rendered":"
Vol 437|20 October 2005|doi:10.1038\/nature04107 LETTERS Adaptive evolution of non-coding DNA in Drosophila Peter Andolfatto1 A large fraction of eukaryotic genomes consists of DNA that is\u2026<\/p>\n","protected":false},"author":3,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[9033,7586,1386,7589,9034,9031,7587,7588,7584,7585,9032],"tags":[8809,7838,8811,8808,8813,8807,8806,8814,8810,8812],"class_list":["post-107273","post","type-post","status-publish","format-standard","hentry","category-assignment-help-uae","category-best-dissertation-writing-service-in-australia","category-best-research-paper-writing-services-in-usa","category-capstone-project-writing-service","category-cheap-essay-writing-services-china","category-essay-writing-service-usa","category-legitimate-phd-dissertation-writing-services-usa","category-term-paper-writing-service","category-thesis-writing-service-in-canada","category-uae-best-dissertation-writing-help-service","category-write-my-essay-australia","tag-best-master-thesis-writing-service","tag-best-thesis-writing-services","tag-cheap-thesis-writing-services","tag-master-thesis-help","tag-master-thesis-writing-service","tag-thesis-help-online-free","tag-thesis-writing-help-online","tag-thesis-writing-services-cost","tag-thesis-writing-services-near-me","tag-thesis-writing-services-reviews"],"_links":{"self":[{"href":"https:\/\/www.studyproessays.com\/help\/wp-json\/wp\/v2\/posts\/107273","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.studyproessays.com\/help\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.studyproessays.com\/help\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.studyproessays.com\/help\/wp-json\/wp\/v2\/users\/3"}],"replies":[{"embeddable":true,"href":"https:\/\/www.studyproessays.com\/help\/wp-json\/wp\/v2\/comments?post=107273"}],"version-history":[{"count":0,"href":"https:\/\/www.studyproessays.com\/help\/wp-json\/wp\/v2\/posts\/107273\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.studyproessays.com\/help\/wp-json\/wp\/v2\/media?parent=107273"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.studyproessays.com\/help\/wp-json\/wp\/v2\/categories?post=107273"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.studyproessays.com\/help\/wp-json\/wp\/v2\/tags?post=107273"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}