The effects of natural selection across molecular pathways in Drosophila melanogaster
© Vedanayagam and Garrigan. 2015
Received: 13 May 2015
Accepted: 30 August 2015
Published: 21 September 2015
Whole-genome RNA interference post-transcriptional silencing (RNAi) is a widely used method for studying the phenotypic effects of knocking down individual genes. In this study, we use a population genomic approach to characterize the rate of evolution for proteins affecting 26 RNAi knockdown phenotypes in Drosophila melanogaster.
We find that only two of the 26 RNAi knockdown phenotypes are enriched for rapidly evolving proteins: innate immunity and regulation of Hedgehog signaling. Among all genes associated with an RNAi knockdown phenotype, we note examples in which the adaptively evolving proteins play a well-defined role in a given molecular pathway. However, most adaptively evolving proteins are found to perform more general cellular functions. When RNAi phenotypes are grouped into categories according to cellular function, we find that genes involved in the greatest number of phenotypic categories are also significantly more likely to have a history of rapid protein evolution.
We show that genes that have been demonstrated to have a measurable effect on multiple molecular phenotypes show higher rates of protein evolution than genes having an effect on a single category of phenotype. Defining pleiotropy in this way yields very different results than previous studies that define pleiotropy by the number of physical interactions, which show highly connected proteins tend to evolve more slowly than lowly connected proteins. We suggest that a high degree of pleiotropy may increase the likelihood of compensatory substitution, consistent with modern theoretical work on adaptation.
KeywordsAdaptation Drosophila melanogaster Pleiotropy RNA interference
The visibility of a protein to natural selection depends upon the phenotypic consequences of mutations to its regulatory and structural sequences. For most proteins, the phenotypic consequences of mutations first manifest at the cellular level, specifically with respect to the protein’s ability to participate in a suite of molecular interactions. This context proximally determines both the level of sequence constraint and how often a protein produces evolutionary adaptations. For over forty years, biologists have endeavored to identify variables that predict the rate of protein evolution [1, 2]. Proteome-level statistical analyses generally find that expression pattern, breadth of interactions, and the genomic context of coding sequences are all correlated with the rate of protein evolution . Even the position of proteins in molecular interaction pathways (upstream or downstream) accounts for some variance in evolutionary rate . It is also widely appreciated that molecular pathways involved in immunity or genome defense are often enriched for adaptively evolving proteins [5, 6]. As functional and genomic data continue to accumulate, the tools are now available to address in detail whether certain categories of pathways are more or less impacted by natural selection.
The targeted knockdown of individual genes with short interfering RNA molecules (RNAi) is routinely used to assay the relative effect of proteins on a measurable phenotype of interest . While the phenotypic effects of gene knockdown are not necessarily representative of the effects of all possible point mutations , they are indicative of the relative importance of the protein in different molecular pathways. This study presents an evolutionary analysis of proteins found to have significant knockdown effects in 26 whole-genome RNAi experiments in Drosophila melanogaster. We ask whether groups of genes affecting a given phenotype are preferentially subject to positive natural selection, relative to a random sample from the genome. Furthermore, we identify which of these genes are most impacted by recurrent positive selection. The results indicate that both immunity and cell signaling pathways are enriched for rapidly evolving proteins and that proteins with wider pleiotropic effects are more rapidly evolving than proteins that affect a narrower range of phenotypes.
Results and discussion
Natural selection across pathways
Each RNAi experiment k yields a set of n k genes that, upon knockdown, cause a significant measurable change to the phenotype. The threshold for statistical significance of the magnitude of change is standardized across studies. Then, using population genomic data from D. melanogaster and two outgroup genomes, for each phenotype, we estimate the direction of selection statistic (DoS), which is defined as the difference between the proportion of substitutions and polymorphisms that are nonsynonymous. Under strictly neutral evolution, DoS is expected to be zero, and it is positive when the proportion of substitutions that are nonsynonymous is higher than the proportion of polymorphisms that are nonsynonymous, indicative of positive natural selection. Alternatively, DoS is negative when the proportion of polymorphisms that are nonsynonymous is higher than the proportion of substitutions that are nonsynonymous, suggestive of weak negative selection . For each set of n k genes influencing a phenotype, we determine the average DoS for all genes associated with that phenotype, and using a two-tailed randomization test, we further determine whether an RNAi phenotype is enriched for genes subject to positive natural selection, compared to a random sample from the genome. We note that this test is designed to detect lineage-specific recurrent natural selection.
The 26 RNAi knockdown phenotypes surveyed in this study. Bolded lines indicate phenotypes that are enriched for proteins that significantly deviate from the genome average in their direction of selection (DoS) statistic
RNAi knockdown phenotype
Regulation of intracellular signal transduction
Akt-TOR signaling decrease
Akt-TOR signaling increase
Hippo signaling decrease
Hippo signaling increase
JAK/STAT signaling decrease
JAK/STAT signaling increase
RTK-Ras-ERK signaling decrease
RTK-Ras-ERK signaling increase
Cell surface receptor signaling pathway
Hedgehog signaling decrease
Hedgehog signaling increase
Notch signaling decrease
Notch signaling increase
Toll signaling decrease
Toll signaling increase
Wnt signaling activity
Regulation of transposon integration
Blood TE activity increase
Burdock TE activity increase
HeTA TE activity increase
TAHRE TE activity increase
Innate immune response
Influenza replication decrease
M. fortuitum infection decrease
Regulation of extent of cell growth
Cell size regulation
Cell growth and viability
Regulation of circadian rhythm
Hypoxia-inducible factor-1alpha signaling pathway
Hypoxia induced transcription
The role of positively selected proteins
The 11 genes experiencing recurrent positive natural selection
RTK-Ras-ERK signaling decrease
RTK-Ras-ERK signaling increase
Decreased cell viability
Blood TE activity
RTK-Ras-ERK signaling decrease
RTK-Ras-ERK signaling decrease
RTK-Ras-ERK signaling increase
Although a handful of adaptively evolving proteins in signaling pathways are exclusive to just one phenotype, many proteins also play a role in multiple cell signaling pathways. For example, the RasGap1 and Dref genes encode proteins with a history of recurrent positive selection and are involved in multiple signaling pathway phenotypes (Table 2). Both RasGap1 and Dref play a role in Ras-mediated signal transduction [24, 25], which activates multiple downstream signaling pathways. Other positively selected proteins influencing cell signaling activity perform more general cellular functions. For instance, two nucleoporin genes (Nup153, Nup205) are both positively selected. While Nup153 is involved in multiple signaling RNAi knockdown phenotypes, Nup205 is identified as significantly influencing the Wnt signaling pathway. Nucleoporin genes encode components of the nuclear pore complex and therefore play a very general role in nuclear transport; these genes have previously been shown to be adaptively evolving in D. melanogaster . Interestingly, one of the knockdown phenotypes influenced by Nup153 is also influenced by the positively selected CCCH-type zinc finger gene ZC3H3; ZC3H3 encodes a necessary component linking mRNA polyadenylation with nuclear export . Both groups of proteins are known to interact with viral proteins [28, 29], which may be a potential source of selective pressure.
In addition to cell signaling pathways, our analysis identifies a new candidate for positively selected proteins in the piRNA pathway. The piRNA pathway generates small RNAs that suppress transposable element (TE) activity in the germline . The piRNA effector proteins Mael, Armi, Aub, and Spn-E have been previously shown to experience positive natural selection in the Drosophila phylogeny . Our analysis identifies a gene pcm, which both affects TE activity and shows an increased rate of adaptive amino acid substitutions in the D. melanogaster lineage (Table 2). pcm encodes a 5′−3′ exoribonuclease that has been previously characterized as having significant sequence conservation between Drosophila, mouse, and Saccharomyces . The Pcm protein is recruited by protein complexes involved in both non-sense mediated mRNA decay (NMD) and RNA interference to degrade targeted mRNAs in cytoplasmic P-bodies .
The effects of pleiotropy
Direction of selection (DoS) statistic for 20 genes involved in three different categories of RNAi knockdown phenotype
Drosophila melanogaster represents one of most mature and powerful systems in genetics and functional genomics and is widely used as a model for studying the genetic basis of human disease [41, 42]. In particular, studies of D. melanogaster have led to significant advances in basic developmental, neurological, and immunological genetics. It is often stated that D. melanogaster is an appropriate genetic model because more than 60 % of the genes found in the D. melanogaster have human homologs  and that genes involved in key developmental pathways are “conserved” and functionally orthologous between humans and flies . For D. melanogaster to be a viable human disease model, it is important to first understand the phenotypic effects of lineage-specific adaptations. While our results recapitulate the well-known conclusion that proteins affecting immunity and genome defense pathways are more likely to fix adaptive mutations, we also find that proteins affecting a suite of cell signaling pathways that are important for metazoan development are also fixing adaptive mutations in the D. melanogaster lineage at a significantly higher rate than the genome average. Our meta-analytical approach is conservative, such that we seek to minimize type I error in a potentially noisy data set. Less stringent criteria for statistical significance may, in fact, yield a different set of conclusions. However, our stringency adds to our confidence that the results do reflect the underlying biological realities concerning the molecular phenotypic effects of adaptive protein evolution.
In general, we refrain from speculating on the nature of the selective pressures driving the inferred adaptive evolution. However, it is important to note that the traditional MK framework used here is designed to detect recurrent bouts of adaptive evolution. One common explanation for recurrent positive selection is conflict due to an ongoing “arms race” between a host genome and either exogenous factors, such as pathogens , or endogenous selfish genetic elements, such as TEs or meiotic drive loci . An “arms race” scenario would certainly apply to proteins involved in immunity or genome defense, as well as to proteins with general functions that interact with exogenous protein , such as is the case with the nucleoporins. Another potential source of recurrent positive selection is compensatory evolution . Compensatory substitutions may resolve any antagonistic effects on fitness caused by an initial adaptive substitution. For instance, if strong positive selection fixes a mutation based on one aspect of the protein’s function, but that mutation also has lesser, deleterious effects on other aspects of the protein’s function, then natural selection will favor subsequent mutations that ameliorate these antagonistic effects. Our inference that proteins affecting a diverse range of molecular pathways are also more likely to experience adaptive evolution is consistent with this hypothesis. This conclusion lends support to two previous results that highlight the potential importance of compensatory evolution. The first is taken from evolutionary theory on the “cost of complexity”, which predicts that adaptive walks are characterized by initial mutations with large fitness effects, followed by mutations of smaller effect . Empirical evidence also suggests that compensatory substitution is common: amino acid substitutions in D. melanogaster are observed to cluster according to their location in a protein’s tertiary structure , suggesting compensatory substitutions occur to preserve functional integrity. Because the MK-based framework is a widely used tool to infer the action of natural selection, the ability to distinguish “arms race” scenarios from compensatory evolution promises to bring unique new insights into the mode of protein evolution.
Data for 26 RNAi screens in Drosophila melanogaster are compiled from the GenomeRNAi database, release 3.0 . All screens report standardized Z scores, which measure the effect that knocking down a single gene has on a phenotype, relative to that of a control gene. Across studies, we consider genes with Z<−3 or Z>3 to have significant effects on a phenotype. Positive and negative tails of Z are sampled depending on the phenotype, for example the negative tail of Z is taken for the JAK/STAT signaling decrease and the positive tail is taken for the JAK/STAT signaling increase. Off-target effects in RNAi screens may potentially overestimate the effects of single genes , all of the RNAi experiments cited in this study report designing dsRNA to be specific to single genes and, in some cases, knockdown effects are further validated by a variety of methods. Individual RNAi phenotypes are grouped into categories that reflect the deepest level of functional ontology that are shared by all of the phenotypes.
Population genomic data
Reference-based genome assemblies of six European and nine sub-Saharan African strains of D. melanogaster (Additional file 1: Table S1) are generated from short-read data in the NCBI short read archive . Reads are mapped to the genome of the reference D. melanogaster strain y 1;c n 1 b w 1 s p 1 (version 5.45) using the BWA software . Variants are called using the POPBAM software with default settings . Gene alignments are then constructed for the longest transcript per gene from the FlyBase mRNA annotations, using the Perl script PBsnp2fa.pl (https://github.com/skingan/PBsnp2fa.pl). A total of 13329 alignments were initially constructed. Ancestral and derived states are inferred by aligning to the genomes of both D. simulans strain MD063  and D. yakuba strain Tai18E2 . Requiring sequence alignment to both D. simulans and D. yakuba limits the data set to 11148 total gene alignments (2839 gene alignments are dropped) and it is likely that very rapidly evolving genes may not appear in the final data set.
Tests of natural selection
To determine the relative effects of natural selection across different RNAi knockdown phenotypes, we perform two analyses. First, we ask whether the genes associated with each RNAi knockdown phenotype, as a group, are enriched for amino acid substitutions (indicating adaptive evolution). We ask whether the genes significantly affecting each phenotype have an increased number of nonsynonymous substitutions compared to nonsynonymous polymorphisms using the direction of selection (DoS) statistic . The DoS statistic is defined as the difference between the proportion of nonsynonymous substitutions (D N ) to the sum of synonymous substitutions (D S ) and nonsynonymous substitutions (D N ) and the proportion of nonsynonymous polymorphisms (P S ) to the sum of synonymous polymorphisms (P S ) and nonsynonymous polymorphisms (P N ), given as DoS=D N /(D S +D N )−P N /(P S +P N ). Statistical significance of DoS is assessed by a bootstrap procedure, in which a null distribution is calculated by selecting a random sample of N genes from the genome, where N is the number of genes in the phenotype to be evaluated. Significance is assessed for DoS using a two-tailed approach, therefore empirical values are considered significant if they fall outside 0.975 quantile of the null distribution. For each phenotype, 10000 bootstrap replicates are performed using the statistical programming language R.
Our second analysis uses single locus MK tests to evaluate individual gene alignments for signatures of positive natural selection. The MK test considers the null hypothesis that the ratio of nonsynonymous (D N ) and synonymous (D S ) substitutions between D. melanogaster and D. simulans is equal to the ratio of nonsynonymous (P N ) and synonymous (P S ) polymorphisms within D. melanogaster . Given that the ratio of P N /P S forms the expectation for the ratio of D N /D S , we calculate probability of obtaining D N higher than the observed value using a one-tailed Fisher’s exact test (FET). Gene alignments with fewer than six sites in any of the marginal counts are considered to have zero power . Of the original 11148 MK tests, by the above criterion, 4063 tests are determined to have zero power and are subsequently removed  leaving 7085 valid tests. It is likely that removing low-power tests results in the elimination of genes experiencing a low rate of neutral mutation, but since there is no predictable relationship between neutral mutation rate and the distribution of fitness effects , there is no concrete a priori reason to believe this procedure will systematically bias our analysis of natural selection. From the remaining valid MK tests, using the Perl script mk-fdr.pl (https://github.com/dgarriga/mk-fdr) the proportion that are truly null is estimated to be 0.978, using a method designed to analyze P-value distributions from conservative tests . At the 5 % level of significance, this corresponds to a false discovery rate of 40.9 %. However, we only consider tests with P FET<0.01 to be statistically significant, which corresponds to a false discovery rate of 22.7 %. Finally, it should be noted that we observe a significant negative correlation between the number of codons in a gene and the MK test P-value (r 2=0.01334; P≪0.001).
We are grateful to Sarah B. Kingan and Amanda Larracuente for insightful discussion and comments on previous drafts of this manuscript. We would also like to thank Anthony J. Geneva, who contributed analyses to previous drafts of this manuscript. This work was made possible by National Institutes of Health grant R01-ODO1054801 to DG and National Science Foundation grant DEB-1209536 to DG and JPV.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
- Kimura M, Ohta T. On some principles governing molecular evolution. Proc Natl Acad Sci USA. 1974; 71:2848–852.PubMed CentralView ArticlePubMedGoogle Scholar
- Wilson AC, Carlson SS, White TJ. Biochemical evolution. Annu Rev Biochem. 1977; 46:573–639.View ArticlePubMedGoogle Scholar
- Pal C, Papp B, Lercher MJ. An integrated view of protein evolution. Nat Rev Genet. 2006; 7:337–48.View ArticlePubMedGoogle Scholar
- Alvarez-Ponce D, Guirao-Rico S, Orengo DJ, Segarra C, Rozas J, Aguadé M. Molecular population genetics of the insulin/TOR signal transduction pathway: a network-level analysis in Drosophila melanogaster. Mol Biol Evol. 2012; 29:123–32.View ArticlePubMedGoogle Scholar
- Sackton TB, Lazzaro BP, Schlenke TA, Evans JD, Hultmark D, Clark AG. Dynamic evolution of the innate immune system in Drosophila. Nat Genet. 2007; 39:1461–1468.View ArticlePubMedGoogle Scholar
- Obbard DJ, Gordon KHJ, Buck AH, Jiggins FM. The evolution of RNAi as a defence against viruses and transposable elements. Philos T Roy Soc B. 2009; 364:99–115.View ArticleGoogle Scholar
- Mohr S, Bakal C, Perrimon N. Genomic screening with RNAi: results and challenges. Ann Rev Biochem. 2010; 79:37–64.PubMed CentralView ArticlePubMedGoogle Scholar
- Stern DL. Perspective: evolutionary developmental biology and the problem of variation. Evolution. 2000; 54:1079–1091.View ArticlePubMedGoogle Scholar
- Stoletzki N, Eyre-Walker A. Estimation of the neutrality index. Mol Biol Evol. 2011; 28:63–70.View ArticlePubMedGoogle Scholar
- Georgel P, Naitza S, Kappler C, Ferrandon D, Zachary D, Swimmer C, et al. Drosophila immune deficiency (IMD) is a death domain protein that activates antibacterial defense and can promote apoptosis. Dev Cell. 2001; 1:503–14.View ArticlePubMedGoogle Scholar
- Briscoe J, Therond PP. The mechanisms of Hedgehog signalling and its roles in development and disease. Nat Rev Mol Cell Biol. 2013; 14:416–29.View ArticlePubMedGoogle Scholar
- Müller B, Basler K. The repressor and activator forms of Cubitus interruptus control Hedgehog target genes through common generic Gli-binding sites. Development. 2000; 127:2999–3007.PubMedGoogle Scholar
- Aikin RA, Ayers KL, Therond PP. The role of kinases in the Hedgehog signalling pathway. EMBO Rep. 2008; 9:330–6.PubMed CentralView ArticlePubMedGoogle Scholar
- Maslov S, Sneppen K, Eriksen KA, Yan KK. Upstream plasticity and downstream robustness in evolution of molecular networks. BMC Evol Biol. 2004; 4:9.PubMed CentralView ArticlePubMedGoogle Scholar
- Langley CH, Stevens K, Cardeno C, Lee YCG, Schrider DR, Pool JE, et al. Genomic variation in natural populations of Drosophila melanogaster. Genetics. 2012; 192:533–98.PubMed CentralView ArticlePubMedGoogle Scholar
- Wehr MC, Holder MV, Gailite I, Saunders RE, Maile TM, Ciirdaeva E, et al. Salt-inducible kinases regulate growth through the Hippo signalling pathway in Drosophila. Nat Cell Biol. 2013; 15:61–71.PubMed CentralView ArticlePubMedGoogle Scholar
- Stabell M, Larsson J, Aalen RB, Lambertsson A. Drosophila dSet2 functions in H3-K36 methylation and is required for development. Biochem Biophys Res Commun. 2007; 359:784–9.View ArticlePubMedGoogle Scholar
- Gerhart J. 1998 warkany lecture: signaling pathways in development. Teratology. 1999; 60:226–39.View ArticlePubMedGoogle Scholar
- Pires-daSilva A, Sommer RJ. The evolution of signalling pathways in animal development. Nat Rev Genet. 2003; 4:39–49.View ArticlePubMedGoogle Scholar
- Serra F, Arbiza L, Dopazo J, Dopazo H. Natural selection on functional modules, a genome-wide analysis. PLoS Comput Biol. 2011; 7:1001093.View ArticleGoogle Scholar
- Keys DN, Lewis DL, Selegue JE, Pearson BJ, Goodrich LV, Johnson RL, et al. Recruitment of a hedgehog regulatory circuit in butterfly eyespot evolution. Science. 1999; 283:532–4.View ArticlePubMedGoogle Scholar
- Kuraku S, Usuda R, Kuratani S. Comprehensive survey of carapacial ridge-specific genes in turtle implies co-option of some regulatory genes in carapace evolution. Evol Dev. 2005; 7:3–17.View ArticlePubMedGoogle Scholar
- Moczek AP, Nagy LM. Diverse developmental mechanisms contribute to different levels of diversity in horned beetles. Evol Dev. 2005; 7:175–85.View ArticlePubMedGoogle Scholar
- Gaul U, Mardon G, Rubin GM. A putative Ras GTPase activating protein acts as a negative regulator of signaling by the Sevenless receptor tyrosine kinase. Cell. 1992; 68:1007–1019.View ArticlePubMedGoogle Scholar
- Huang AM, Rubin GM. A misexpression screen identifies genes that can modulate RAS1 pathway signaling in Drosophila melanogaster. Genetics. 2000; 156:1219–1230.PubMed CentralPubMedGoogle Scholar
- Presgraves DC, Stephan W. Pervasive adaptive evolution among interactors of the Drosophila hybrid inviability gene, Nup96. Mol Biol Evol. 2007; 24:306–14.View ArticlePubMedGoogle Scholar
- Hurt JA, Obar RA, Zhai B, Farny NG, Gygi SP, Silver PA. A conserved CCCH-type zinc finger protein regulates mRNA nuclear adenylation and export. J Cell Biol. 2009; 185:265–77.PubMed CentralView ArticlePubMedGoogle Scholar
- Varadarajan P, Mahalingam S, Liu P, Ng SB, Gandotra S, Dorairajoo DS, et al. The functionally conserved nucleoporins Nup124p from fission yeast and the human Nup153 mediate nuclear import and activity of the Tf1 retrotransposon and HIV-1 Vpr. Mol Biol Cell. 2005; 16:1823–1838.PubMed CentralView ArticlePubMedGoogle Scholar
- Kozaki T, Takahama M, Misawa T, Matsuura Y, Akira S, Saitoh T. Role of zinc-finger anti-viral protein in host defense against Sindbis virus. Int Immunol. 2015; 27:357–64.View ArticlePubMedGoogle Scholar
- Aravin AA, Sachidanandam R, Girard A, Fejes-Toth K, Hannon GJ. Developmentally regulated piRNA clusters implicate MILI in transposon control. Science. 2007; 316:744–7.View ArticlePubMedGoogle Scholar
- Till DD, Linz B, Seago JE, Elgar SJ, Marujo PE, de Lourdes Elias M, et al. Identification and developmental expression of a 5’-3’ exoribonuclease from Drosophila melanogaster. Mech Develop. 1998; 79:51–5.View ArticleGoogle Scholar
- Rehwinkel J, Behm-Ansmant I, Gatfield D, Izaurralde E. A crucial role for GW182 and the DCP1:DCP2 decapping complex in miRNA-mediated gene silencing. RNA. 2005; 11:1640–1647.PubMed CentralView ArticlePubMedGoogle Scholar
- Hahn MW, Kern AD. Comparative genomics of centrality and essentiality in three eukaryotic protein-interaction networks. Mol Biol Evol. 2005; 22:803–6.View ArticlePubMedGoogle Scholar
- Luisi P, Alvarez-Ponce D, Pybus M, Fares MA, Bertranpetit J, Laayouni H. Recent positive selection has acted on genes encoding proteins with more interactions within the whole human interactome. Genome Biol Evol. 2015; 7:1141–1154.PubMed CentralView ArticlePubMedGoogle Scholar
- Andolfatto P. Controlling type-I error of the McDonald-Kreitman test in genomewide scans for selection on noncoding DNA. Genetics. 2008; 180:1767–1771.PubMed CentralView ArticlePubMedGoogle Scholar
- Sundaram MV. The love-hate relationship between Ras and Notch. Gene Dev. 2005; 19:1825–1839.View ArticlePubMedGoogle Scholar
- Ragab A, Buechling T, Gesellchen V, Spirohn K, Boettcher AL, Boutros M. Drosophila Ras/MAPK signalling regulates innate immune responses in immune and intestinal stem cells. EMBO J. 2011; 30:1123–1136.PubMed CentralView ArticlePubMedGoogle Scholar
- Heallen T, Zhang M, Wang J, Bonilla-Claudio M, Klysik E, Johnson RL, et al. Hippo pathway inhibits Wnt signaling to restrain cardiomyocyte proliferation and heart size. Science. 2011; 332:458–61.PubMed CentralView ArticlePubMedGoogle Scholar
- di Magliano MP, Sekine S, Ermilov A, Ferris J, Dlugosz AA, Hebrok M. Hedgehog/Ras interactions regulate early stages of pancreatic cancer. Gene Dev. 2006; 20(22):3161–173.View ArticleGoogle Scholar
- Lauth M. RAS and Hedgehog– partners in crime. Front Biosci (Landmark Ed). 2011; 16:2259–270.View ArticleGoogle Scholar
- Bier E. Drosophila, the golden bug, emerges as a tool for human genetics. Nat Rev Genet. 2005; 6:9–23.View ArticlePubMedGoogle Scholar
- Wangler MF, Yamamoto S, Bellen HJ. Fruit flies in biomedical research. Genetics. 2015; 199(3):639–53.View ArticlePubMedGoogle Scholar
- Daugherty MD, Malik HS. Rules of engagement: molecular insights from host-virus arms races. Ann Rev Genet. 2012; 46:677–700.View ArticlePubMedGoogle Scholar
- Werren JH. Selfish genetic elements, genetic conflict, and evolutionary innovation. Proc Natl Acad Sci USA. 2011; 108:10863–10870.PubMed CentralView ArticlePubMedGoogle Scholar
- Daub JT, Hofer T, Cutivet E, Dupanloup I, Quintana-Murci L, Robinson-Rechavi M, et al. Evidence for polygenic adaptation to pathogens in the human genome. Mol Biol Evol. 2013; 30(7):1544–1558.View ArticlePubMedGoogle Scholar
- Akashi H, Osada N, Ohta T. Weak selection and protein evolution. Genetics. 2012; 192:15–31.PubMed CentralView ArticlePubMedGoogle Scholar
- Orr HA. The population genetics of adaptation: the adaptation of DNA sequences. Evolution. 2002; 56:1317–1330.View ArticlePubMedGoogle Scholar
- Callahan B, Neher RA, Bachtrog D, Andolfatto P, Shraiman BI. Correlated evolution of nearby residues in Drosophilid proteins. PLoS Genet. 2011; 7:1001315.View ArticleGoogle Scholar
- Schmidt EE, Pelz O, Buhlmann S, Kerr G, Horn T, Boutros M. GenomeRNAi: a database for cell-based and in vivo RNAi phenotypes, 2013 update. Nucleic Acids Res. 2013; 41:1021–1026.View ArticleGoogle Scholar
- Ma Y, Creanga A, Lum L, Beachy PA. Prevalence of off-target effects in Drosophila RNA interference screens. Nature. 2006; 443:359–63.View ArticlePubMedGoogle Scholar
- Pool J, Corbett-Detig R, Sugino R, Stevens K, Cardeno C, Crepeau M, et al. Population genomics of sub-Saharan Drosophila melanogaster: African diversity and non-African admixture. PLoS Genet. 2012; 8:1003080.View ArticleGoogle Scholar
- Li H, Durbin R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics. 2009; 25:1754–1760.PubMed CentralView ArticlePubMedGoogle Scholar
- Garrigan D. POPBAM: tools for evolutionary analysis of short read sequence alignments. Evol Bioinform. 2013; 9:343–53.View ArticleGoogle Scholar
- Garrigan D, Kingan SB, Geneva AJ, Andolfatto P, Clark AG, Thornton KR, et al. Genome sequencing reveals complex speciation in the Drosophila simulans clade. Genome Res. 2012; 22:1499–1511.PubMed CentralView ArticlePubMedGoogle Scholar
- Clark AG, Eisen MB, Smith DR, Bergman CM, Oliver B, Markow TA, et al. Evolution of genes and genomes on the Drosophila phylogeny. Nature. 2007; 450:203–18.View ArticlePubMedGoogle Scholar
- McDonald JH, Kreitman M. Adaptive protein evolution at the Adh locus in Drosophila. Nature. 1991; 351:652–4.View ArticlePubMedGoogle Scholar
- Dialsingh I, Austin SR, Altman NS. Estimating the proportion of true null hypotheses when the statistics are discrete. Bioinformatics. 2015; 31:2303–309.View ArticlePubMedGoogle Scholar
- Eyre-Walker A, Keightley PD. The distribution of fitness effects of new mutations. Nat Rev Genet. 2007; 8:610–8.View ArticlePubMedGoogle Scholar