|

 

Op-gbev120037 1046.1053

Weak 50-mRNA Secondary Structures in Short EukaryoticGenes Yang Ding, Premal Shah, and Joshua B. Plotkin* Department of Biology, University of Pennsylvania *Corresponding author: E-mail: [email protected].
Accepted: September 21, 2012 Experimental studies of translation have found that short genes tend to exhibit greater densities of ribosomes than long genes ineukaryotic species. It remains an open question whether the elevated ribosome density on short genes is due to faster initiation orslower elongation dynamics. Here, we address this question computationally using 50-mRNA folding energy as a proxy for translationinitiation rates and codon bias as a proxy for elongation rates. We report a significant trend toward reduced 50-secondary structure inshorter coding sequences, suggesting that short genes initiate faster during translation. We also find a trend toward higher 50-codonbias in short genes, suggesting that short genes elongate faster than long genes. Both of these trends hold across a diverse set ofeukaryotic taxa. Thus, the elevated ribosome density on short eukaryotic genes is likely caused by differential rates of initiation, ratherthan differential rates of elongation.
Key words: translation initiation, ribosome density, codon bias, gene length.
Selection for translational efficiency remains the dominant explanation for systematic variation in codon usage among Synonymous sites in coding sequences have long been used as the genes in a genome, in diverse taxa a neutral yardstick against which to compare amino acid chan- In accordance with this explanation, codon bias toward ging substitutions, in the hope of detecting either purifying or the most abundant iso-accepting tRNA species is generally positive selection on proteins ( strongest in those genes expressed at high levels, where effi- ciency would confer the greatest selective benefit to the cell.
Nonetheless, synonymous mutations are known to Nonetheless, the specific mechanisms by which codon bias experience selection in many cases confers relative fitness gains are actively debated Our understanding of the dynamics of gene translation, for a variety of mechanisms, including the efficiency of and the role of codon bias in translation, will benefit from gene translation, the stability of mRNAs ; new experimental techniques that parse the detailed kinetics of translation across the entire transcriptome. Especially pro- mising are techniques that use high-throughput sequencing cially near the translation initiation site of ribosome-protected RNA to determine a "ribosomal foot- ; and the regulation of splicing, print" on each mRNA (, ; among others ). The fact that syn- onymous mutations have phenotypic and fitness conse- ) with greater accuracy quences complicates the interpretation of measures of than earlier, polysome-based techniques ).
selection, such as the ratio of substitution rates at synonymous Among many other intriguing findings, these experiments and nonsynonymous sites, dN/dS have shown that the cell-wide average profile of ribosome densities in yeast exhibits a trend of decreasing ribosome dens- ity with codon position, from 50 to 30—an observation that has ß The Author(s) 2012. Published by Oxford University Press on behalf of the Society for Molecular Biology and Evolution.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by-nc/3.0/), which permits non-commercialreuse, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact [email protected].
Genome Biol. Evol. 4(10):1046–1053.
Advance Access publication October 3, 2012 Weak 50-mRNA Secondary Structures in Short Eukaryotic Genes been explained, in part, by a trend toward less biased codon as well as the relationship between ORF length and 50-codon usage in the 50-ends of genes, associated presumably with bias. As described earlier, we use these two measures as slower elongation and thus higher ribosome density ( proxies for the initiation rates and early elongation rates of genes. In particular, for each C. elegans transcript, we com- Aside from the 50-ramp of elevated ribosome densities, puted its predicted folding energy from nucleotide 4 to +37 sequencing () and polysome gradients in relative to start, using RNAfold ( budding yeast have also revealed another, and we computed the CAI of its first 50 codons.
possibly independent finding: shorter mRNAs tend to have a (We systematically explore alternative definitions of 50-CAI greater overall density of ribosomes than longer mRNAs. The same trend has been found in mouse, human, fruit fly, We performed a Spearman rank correlation test between Arabidopsis, malaria, and fission yeast: shorter Open 50-mRNA folding energy and ORF length, among the 29,857 Reading Frames (ORFs) tend to exhibit more densely packed transcripts in C. elegans (Assembly WS220). We similarly per- formed a rank correlation test between 50-CAI values and ORF lengths. Our expectation was that compared with long genes, short genes should tend to have faster initiation rates and/or debate about the cause of this trend. Some authors have slower early elongation rates—to explain the tendency toward attributed this relationship to a constant-length ramp of ele- elevated ribosome densities on short genes vated 50-density on all transcripts due to elongation dynamics ) (so that shorter transcripts would be observed to have larger overall ribosome density); and others have attributed this trend to an increased rate of initi- alternative mechanisms, we might in principal expect the ation in short yeast genes causing an increased density of initiation-driven mechanism to be a stronger determinant of ribosome densities ( a result, at present, it is unclear whether the greater overall density of ribosomes on short yeast genes is caused by a In accordance with these expectations, we found a greater rate of initiation for such genes or a slower rate of significant negative rank correlation (Spearman rho ¼ 0.12, early elongation in those genes.
P < 7  1090) between 50-mRNA folding energy and ORF Against this backdrop of open questions, here we analyze length, indicating a tendency toward weaker mRNA structure the relationship between ORF length and measures of initi- and presumably faster initiation in short C. elegans genes ation and early elongation rates, across a diverse set of eu- ). On the other hand, we also found a significant nega- karyotic species. As a proxy for the initiation rate of a gene, we tive rank correlation (Spearman rho ¼ 0.16, P < 5  10179) use the computationally predicted energy of its 50-mRNA between 50-CAI and length, suggesting shorter genes tend to structure—a quantity that has been shown experimentally have faster early elongate rates ). Given that shorter to correlate strongly with protein levels () genes have higher CAI and hence faster elongation rates, and which has been subject to natural selection in virtually we would expect a lower ribosomal density for shorter all free-living (; genes contrary to the observed patterns. As a result, we con- and many viral species clude that higher ribosomal densities of shorter genes are As a proxy for the early elongation rate of a gene, most likely explained by faster initiation rates as shown by we use the codon adaptation index (CAI) weaker 50-mRNA secondary structures.
of its early codons In general, byperforming these analyses, we seek to understand whetherthe trend toward elevated ribosome densities in short genes Codon Bias, mRNA Structure, and ORF Length in 120 Eukaryotic Species Given our results in C. elegans, we then asked how broadly faster initiation in those genes, slower early elongation in these trends in gene length and 50-mRNA structure hold those genes, or both.
across eukaryotes. We repeated the 50-mRNA folding energycalculations in 120 eukaryote species and the 50-CAI calcula- tions in 89 of those species for which a reliable reference set ofgenes was available for computing CAI. (The sets of species Codon Bias, mRNA Structure, and ORF Length in used in 50-mRNA folding energy and 50-CAI calculations are Caenorhabditis elegans We first investigated the relationship between ORF length and online). The results of these calculations and their correlations 50-mRNA folding in the model species Caenorhabditis elegans, with ORF length are summarized in Genome Biol. Evol. 4(10):1046–1053.
Advance Access publication October 3, 2012

FIG. 1.—Short C. elegans genes have higher 50-mRNA folding ener- FIG. 2.—Short C. elegans genes have higher 50-CAIs than long gies than long C. elegans genes, suggesting faster initiation in short genes.
C. elegans genes, suggesting faster elongation in short genes. Genes Genes have been binned according to their log (ORF length), with dots have been binned according to their log (ORF length), with dots showing showing the mean computed 50-mRNA folding energy in each bin and the mean computed 50-CAI in each bin and lines showing ±1 standard lines showing ±1 standard deviation. The solid line shows best-fit regres- deviation. The solid line shows best-fit regression (Spearman rho ¼ 0.16, sion (Spearman rho ¼ 0.12, P < 7  1090).
P < 5  10179).
summarizes the proportion of species tested that opposed to long genes; and we argued that the resulting exhibit a negative rank correlation between 50-mRNA folding increase in initiation rates is responsible for the greater density energy and ORF length or between 50-CAI and ORF length. In of ribosomes typically found in short eukaryotic genes.
addition, we report the proportion of species that feature a Nonetheless, we have also found a trend toward increased significant negative correlation, at the 5% significance level.
CAI in the same region, in short genes—and so the possibility As summarized in , the results found in C. elegans hold remains that some subtle patterns of 50-CAI might be respon- very broadly across eukaryotes: approximately 80% of tested sible for the trend observed in mRNA structure. To resolve this eukaryotes exhibit negative correlations between mRNA fold- issue, we have performed a randomization procedure that ing and length and between 50-CAI and length. The prepon- isolates the effects of synonymous codons on 50-mRNA struc- derance of significant negative correlations with ORF length ture, controlling for 50-CAI.
among eukaryotes is itself highly significant, for both 50-mRNA For each species, we randomly shuffled the first 50 folding energy (binomial P < 1011) and 50-CAI (binomial codons of each coding sequence, and we repeated this pro- P < 109)—suggesting a systematic eukaryotic trend toward cess 100 times for each gene. In each such permutation, the faster translation initiation and faster early elongation in short 50-CAI of the gene is preserved, whereas the mRNA structure versus long genes. Thus, our results suggest that the higher is possibly perturbed. We then computed the quantile of the ribosome density observed in shorter eukaryotes genes is likely 50-mRNA folding energy for the true gene sequence with due to faster initiation rates in shorter genes.
respect to this null distribution of permuted sequences.
The distribution of correlations for energy and CAI are pre- Because our hypothesis is that shorter genes are under se- sented in and , and the complete results for each lection for weaker 50-mRNA folding (i.e., higher energy) re- species used in the energy and CAI calculations are presented gardless of 50-CAI, we expect a higher quantile for shorter genes. We tested this expectation by computing the Spearman rank correlation between the length of eachORF in the genome and the quantile of its true mRNA foldingenergy compared with the null distribution.
Weak 50-mRNA Folding in Short Genes, Controlling for As listed in , we observed a negative rank correlation between the energy quantile and the ORF length in the great In the previous sections, we have established a systematic majority species (binomial P value < 6  1015)—indicating trend toward weaker 50-mRNA structure in short genes, as that the trend toward weak mRNA structure in short genes Genome Biol. Evol. 4(10):1046–1053.
Advance Access publication October 3, 2012

Weak 50-mRNA Secondary Structures in Short Eukaryotic Genes Table 1Most Eukaryotic Species Show a Tendency Toward Weak 50-mRNA Structure and High 50-Codon Bias in Shorter Genes Correlations with ORF Length % Species with negative correlation % Species with significant negative correlation % Species with positive correlation % Species with significant positive correlation Two-sided binomial P value NOTE.—In particular, there is a negative rank correlation between 50-mRNA folding energy and ORF length in 82% of the 120 eukaryotic species tested, and similarly, a negative rank correlation between 50-CAI and -ORF length in 83% of the 89 species tested.
The overall tendency toward negative correlations is highly significant, in both cases.
FIG. 3.—The distribution of Spearman rank correlation coefficients FIG. 4.—The distribution of Spearman rank correlation coefficients between 50-energy and -ORF length in 120 eukaryotic species.
between 50-CAI and ORF length in 89 eukaryotic species.
holds even after controlling for 50-CAI. These analyses sub- ribosomes, leading to higher ribosomal densities irrespective stantiate our hypothesis that shorter eukaryotic genes are of the codon composition in the 50 region. As a result, we also under selection to have faster translation initiation rates, verified the robustness of our results by considering the CAI of achieved through weaker 50-mRNA folding.
entire ORF, producing the same qualitative, but slightlyweaker, result (36% positive correlations, 64% negative cor- Robustness of Results relations, two-sided Binomial P value < 0.011. For the com- In the preceding analyses, we calculated 50-CAI using the first plete tabulation of these results see 50 codons of each ORF. We chose this region to coincide as much as possible with the ramp of slow codons reported by Another potential concern that may arise from our 50-CAI We repeated the 50-CAI calculations using calculation is that we excluded sequences shorter than the first 13, 15, 20, 30, 40, and 60 codons and obtained 51 codons. Is it possible that the sequences shorter than similar qualitative results in each case 51 codons could have a different CAI pattern and somehow , online). The ribosomal density diluted the observed CAI pattern? To answer this question, we on a gene might be affected by codons beyond the 50 modified the definition of 50-CAI to include coding sequences region of gene as well. For instance, slow codons in the shorter than 51 codons long, by computing the geometric middle or end of a gene might cause a bottleneck for mean of the relative adaptiveness of all the nonstop codons Genome Biol. Evol. 4(10):1046–1053.
Advance Access publication October 3, 2012 dynamics, modulated by codon bias. This view is in contrast Most Species Exhibit a Tendency Toward Weak 50 Free Energy in with other studies that propose a dominant role of codon Short Genes, Even After Controlling for 50-CAI usage in shaping ribosomal occupancies ().
Correlation between ORF Length Nonetheless, our results do not directly contradict those of and Quantile of Observed 50 Free , however, because those authors con- sidered relative codon usage within each ORF, whereas we Negative correlation have studied absolute codon usage across different ORFs.
Significant negative correlation Other factors such as protein folding Positive correlation and sequence similarity to ribosome binding sites Significant positive correlation may also influence ribosome density. However, One-sided binomial P value such effects are generally not considered as major determin-ants in shaping overall ribosome density ( NOTE.—In the majority of species tested, we find a negative rank correlation between ORF length and the quantile of the observed 50-mRNA free energy These factors, which are difficult to among the free energies of permuted sequences that retain the same 50-CAI quantify systematically, are probably less likely to show sys- value. The tendency toward negative correlations across species is highlysignificant.
tematic trends with respect to ORF length, such as those wehave observed for 50-CAI and 50-mRNA secondary structure.
in the sequence. Again, this did not change our qualitative It is interesting to ask whether there are any commonal- ities among the 22 "counterexample" species in which we observed a positive rank correlation between 50-energy andORF length. What differentiates these organisms from the other eukaryotes we have studied? To answer this question,we examined the phylogenetic relationship of all the studied We have reported a strong trend toward weaker 50-mRNA species and the distribution along this phylogeny of those structure in short genes, when compared with long genes, 22 species exhibiting a positive rank correlation between among eukaryotic species. Moreover, we also observed a ORF length and 50 free energy ( trend toward higher 50-codon bias in short versus long online). Although a few of these genes—indicating that elongation dynamics driven by codon counter examples are clearly closely related sister species, bias is unlikely to be the cause of higher ribosomal densities on overall these 22 species are distributed relatively uniformly short genes. For each individual species, the correlation be- among eukaryotes, as opposed to being mostly monophy- tween ORF length and 50-mRNA folding energy/50-CAI is usu- letic. And so we do not find any obvious commonality ally statistically significant but not strong. Nonetheless, the among these species with respect to their evolutionary his- trend of reduced 50-secondary structure in short coding se- tory and, likely, ecological contexts.
quences was observed in the majority of eukaryotic species Our results on systematically weaker 50-mRNA structure in (82%) tested. The statistical significance of this trend is extra- short genes beg the question: why should short genes experi- ordinarily strong and so too is the biological significance: more ence selection for fast translation initiation? It has been sug- than three-quarters of eukaryotic species exhibit reduced gested that highly expressed genes are shorter in many 50-mRNA structure in short genes.
To the extent that 50-mRNA structure modulates initiation short genes are enriched for constitutively expressed house- keeping and ribosomal genes ), ), our results suggest that faster initiation is respon- which must produce protein as rapidly as possible. This sible for the empirical observation in diverse eukaryotes alone might explain why short genes experience selection for faster initiation ). In addition, house- keeping genes tend to have shorter 50-untranslated regions ) that short mRNAs are more (UTRs) and are under weaker post-transcriptional regulation densely packed with ribosomes than long mRNAs.
Our analyses across a diverse set of eukaryotic species The probability of successful ribosomal binding and substantiates several authors' interpretation of patterns of scanning on an mRNA may depend on the length of its ribosomal densities and ORF length, which have been attrib- 50-UTRs. As a result, genes that require post-transcriptional uted to initiation-driven mechanisms as opposed to elong- regulation tend to have longer 50-UTRs, leading to lower ini- ation effects ).
tiation probabilities ).
Our results confirm that the effects of initiation, modulated In summary, we find that shorter genes have higher by ribosomal binding to the 50-end of mRNA and scanning to 50-mRNA folding energies and codon bias, suggesting that start codon, strongly outweigh those of elongation shorter genes both initiate and elongate faster than longer Genome Biol. Evol. 4(10):1046–1053.
Advance Access publication October 3, 2012 Weak 50-mRNA Secondary Structures in Short Eukaryotic Genes genes. Both of these trends hold across a diverse set of eu- P value using the function spearmanr in the SciPy karyotic taxa. Because faster elongation leads to lower ribo- ) package of Python some densities, the elevated ribosome densities of short We chose 0.05 as the significance level.
eukaryotic genes is a result of initiation rates, rather than We then counted the number of species in which the 50 elongation rates.
free energy has a negative Spearman rank correlation withsequence length and also the number of species in which Materials and Methods the correlations are significant. We calculated a two-tailed Pvalue to assess whether there is an overall trend in the direc- tion of rank correlation between 50-mRNA folding energy and Coding sequences with 4-bp upstream data for most species coding sequence length.
were downloaded from ensembl genomes servers last accessed March 25, 2011).
Calculating 50-CAI The coding sequences of Yarrowia lipolytica with 1,000 bp To obtain an estimate of the translation early elongation rates, upstream sequences and 300 bp downstream sequences we calculated the CAI for the first 50 were downloaded from Ge´nolevures ) codons of each gene. The 50-CAI of a gene is defined as the last accessed March 25, geometric mean of the relative adaptiveness values of all the 2011). All the coding sequences were preprocessed, so that considered codons in a particular gene. The relative adaptive- sequences whose length is not a multiple of 3, those with ness values of each codon are defined as ratio of occurrences premature stop codons, or a continuous string of more than of the codon to occurrences of the most abundant synonym- three ambiguous "N" symbols are discarded. We only con- ous codon, using the ribosomal gene sequences from each sidered coding sequences at least 42 nucleotides long. The species. In the above calculations, we removed coding se- complete list of species used in this study is listed in quences less than 51 codons long. Alternatively, for these short sequences, we also calculated 50-CAI using the whole We identified ribosomal genes for the purpose of comput- sequence and obtained the same qualitative results ing CAI from one of three sources: 1) the ribosomal gene sequences for 24 species were downloaded from theHOGENOMDNA ) database , last accessed Supplementary Material February 1, 2011).
and are available at Orthologous groups of ribosomal genes from the Genome Biology and Evolution online ( HOGENOM database are listed in , online. 2) The ribosomal genesfor 64 species were obtained from Orthologous MAtrixProject , last accessed March 25, 2011). We used Saccharomyces cerevisiae The authors thank two anonymous referees for constructive as our genome of reference and obtained orthologs of its comments. This work was supported by the Burroughs ribosomal genes. The OMA orthologous groups andorganism-specific ribosomal genes are listed in Wellcome Fund, the David and Lucile Packard Foundation, , online. 3) The ribosomal the James S. McDonnell Foundation, the Alfred P. Sloan genes for Y. lipolytica were obtained by performing a protein Foundation, and grant D12AP00025 from the U.S.
blast search against the ribosomal gene coding sequences for Department of the Interior and Defense Advanced Research S. cerevisiae and taking the top hit for each gene provided it Projects Agency to J.B.P. and by the Penn Genome Frontiers Institute to Y.D.
<105. The number of identified ribosomal genes per species in our data set ranged from 19 to 184genes with a median value of 44.
Literature CitedAltenhoff AM, Schneider A, Gonnet GH, Dessimoz C. 2011. OMA 2011: Calculating 50-mRNA Folding Free Energy orthology inference among 1000 complete genomes. Nucleic AcidsRes. 39:D289–D294.
To get an estimate of the translation initiation rates, we used Andersson SG, Kurland CG. 1990. Codon preferences in free-living micro- the program RNAfold from Vienna RNA package organisms. Microbiol Mol Biol Rev. 54:198–210.
to calculate the mRNA folding energy from base Arava Y, Boas FE, Brown PO, Herschlag D. 2005. Dissecting eukaryotic translation and its control by ribosome density mapping. Nucleic Acids 4 to 37 for each gene. For each species, we calculated the 50-folding energy and length of every gene and then obtained Arava Y, et al. 2003. Genome-wide analysis of mRNA translation profiles in the Spearman rank correlation coefficient and a two-tailed Saccharomyces cerevisiae. Proc Natl Acad Sci U S A. 100:3889–3894.
Genome Biol. Evol. 4(10):1046–1053.
Advance Access publication October 3, 2012 Bazzini AA, Lee MT, Giraldez AJ. 2012. Ribosome profiling shows that Hofacker IL, et al. 1994. Fast folding and comparison of RNA secondary miR-430 reduces translation before causing mRNA decay in zebrafish.
structures. Monatshefte Chem. 125:167–188.
Hurowitz EH, Brown PO. 2003. Genome-wide analysis of mRNA lengths in Bettany AJ, et al. 1989. 50-secondary structure formation, in contrast to a Saccharomyces cerevisiae. Genome Biol. 5:R2.
short string of non-preferred codons, inhibits the translation of the Ingolia NT, Ghaemmaghami S, Newman JR, Weissman JS. 2009.
pyruvate kinase mRNA in yeast. Yeast 5:187–198.
Genome-wide analysis in vivo of translation with nucleotide resolution Branco-Price C, Kawaguchi R, Ferreira RB, Bailey-Serres J. 2005.
using ribosome profiling. Science 324:218–223.
Genome-wide analysis of transcript abundance and translation in Ingolia NT, Lareau LF, Weissman JS. 2011. Ribosome profiling of mouse Arabidopsis seedlings subjected to oxygen deprivation. Ann Bot. 96: embryonic stem cells reveals the complexity and dynamics of mam- malian proteomes. Cell 147:789–802.
Brar GA, et al. 2012. High-resolution view of the yeast meiotic program Jones E, Oliphant T, Pearu P. 2001. SciPy: open source scientific tools for revealed by ribosome profiling. Science 335:552–557.
Bulmer M. 1991. The selection-mutation-drift theory of synonymous Keller TE, Mis SD, Jia KE, Wilke CO. 2012. Reduced mRNA codon usage. Genetics 129:897–907.
secondary-structure stability near the start codon indicates functional Capon F, et al. 2004. A synonymous SNP of the corneodesmosin gene genes in prokaryotes. Genome Biol Evol. 4:80–88.
leads to increased mRNA stability and demonstrates association with Kimchi-Sarfaty C, et al. 2007. A "silent" polymorphism in the MDR1 gene psoriasis across diverse ethnic groups. Hum Mol Genet. 13: changes substrate specificity. Science 315:525–528.
Kimura M. 1977. Preponderance of synonymous changes as evi- Cataldo L, Mastrangelo MA, Kleene KC. 1999. A quantitative sucrose dence for the neutral theory of molecular evolution. Nature 267: gradient analysis of the translational activity of 18 mRNA species in testes from adult mice. Mol Hum Reprod. 5:206–213.
Kudla G, Murray AW, Tollervey D, Plotkin JB. 2009. Coding-sequence Chamary JV, Hurst LD. 2005. Evidence for selection on synonymous mu- determinants of gene expression in Escherichia coli. Science 324: tations affecting stability of mRNA secondary structure in mammals.
Genome Biol. 6:R75.
Lackner DH, et al. 2007. A network of multiple regulatory layers shapes Chamary JV, Parmley JL, Hurst LD. 2006. Hearing silence: non-neutral gene expression in fission yeast. Mol Cell. 26:145–155.
evolution at synonymous sites in mammals. Nat Rev Genet. 7:98–108.
Lacsina JR, LaMonte G, Nicchitta CV, Chi JT. 2011. Polysome profiling of David L, et al. 2006. A high-resolution map of transcription in the yeast the malaria parasite plasmodium falciparum. Mol Biochem Parasitol.
genome. Proc Natl Acad Sci U S A. 103:5320–5325.
de Smit MH, van Duin J. 1990. Secondary structure of the ribosome bind- Li GW, Oh E, Weissman JS. 2012. The anti-shine-dalgarno sequence drives ing site determines translational efficiency: a quantitative analysis. Proc translational pausing and codon choice in bacteria. Nature 484: Natl Acad Sci U S A. 87:7668–7672.
Duan J, et al. 2003. Synonymous mutations in the human dopamine re- Lin Z, Li WH. 2012. Evolution of 50 untranslated region length and gene ceptor D2 (DRD2) affect mRNA stability and synthesis of the receptor.
expression reprogramming in yeasts. Mol Biol Evol. 29:81–89.
Hum Mol Genet. 12:205–216.
McDonald JH, Kreitman M. 1991. Adaptive protein evolution at the Adh Duret L. 2002. Evolution of synonymous codon usage in metazoans. Curr locus in Drosophila. Nature 351:652–654.
Opin Genet Dev. 12:640–649.
Muse SV, Gaut BS. 1994. A likelihood approach for comparing Duret L, Mouchiroud D. 1999. Expression pattern and, surprisingly, gene synonymous and nonsynonymous nucleotide substitution rates, length shape codon usage in Caenorhabditis, Drosophila, and with application to the chloroplast genome. Mol Biol Evol. 11: Arabidopsis. Proc Natl Acad Sci U S A. 96:4482–4487.
Eisenberg E, Levanon EY. 2003. Human housekeeping genes are compact.
Oh E, et al. 2011. Selective ribosome profiling reveals the cotranslational Trends Genet. 19:362–365.
chaperone action of trigger factor in vivo. Cell 147:1295–1308.
Eyre-Walker A. 1996. Synonymous codon bias is related to gene length in Penel S, et al. 2009. Databases of homologous gene families for compara- Escherichia coli: selection for translational accuracy? Mol Biol Evol. 13: tive genomics. BMC Bioinformatics 10(Suppl 6):S3.
Plotkin JB, Kudla G. 2011. Synonymous but not the same: the causes and Eyre-Walker A, Bulmer M. 1993. Reduced synonymous substitution rate consequences of codon bias. Nat Rev Genet. 12:32–42.
at the start of enterobacterial genes. Nucleic Acids Res. 21: Qin X, Ahn S, Speed TP, Rubin GM. 2007. Global analyses of mRNA Goldman N, Yang Z. 1994. A codon-based model of nucleotide Genome Biol. 8:R63.
substitution for protein-coding DNA sequences. Mol Biol Evol. 11: Rao YS, et al. 2010. Selection for the compactness of highly expressed genes in Gallus gallus. Biol Direct. 5:35.
Gu W, Zhou T, Wilke CO. 2010. A universal trend of reduced mRNA stability near the translation-initiation site in prokaryotes and eukary- reticulum-bound ribosomes in cellular translation identified by ribo- otes. PLoS Comput Biol. 6:e1000664.
some profiling. J Biol Chem. 287:5518–5527.
Guo H, Ingolia NT, Weissman JS, Bartel DP. 2010. Mammalian microRNAs Reuveni S, Meilijson I, Kupiec M, Ruppin E, Tuller T. 2011. Genome-scale predominantly act to decrease target mRNA levels. Nature 466: analysis of translation elongation with a ribosome flow model. PLoS Comput Biol. 7:e1002127.
Hendrickson DG, et al. 2009. Concordant regulation of translation and Sawyer SA, Hartl DL. 1992. Population genetics of polymorphism and mRNA abundance for hundreds of targets of a human microRNA.
divergence. Genetics 132:1161–1176.
PLoS Biol. 7:e1000238.
Shah P, Gilchrist MA. 2010. Effect of correlated tRNA abundances on Hershberg R, Petrov DA. 2008. Selection on codon bias. Annu Rev Genet.
translation errors and evolution of codon usage bias. PLoS Genet. 6: Hirsh AE, Fraser HB, Wall DP. 2005. Adjusting for selection on synonym- Shah P, Gilchrist MA. 2011. Explaining complex codon usage patterns with ous sites in estimates of evolutionary distance. Mol Biol Evol. 22: selection for translational efficiency, mutation bias, and genetic drift.
Proc Natl Acad Sci U S A. 108:10231–10236.
Genome Biol. Evol. 4(10):1046–1053.
Advance Access publication October 3, 2012 Weak 50-mRNA Secondary Structures in Short Eukaryotic Genes Sharp PM, Averof M, Lloyd AT, Matassi G, Peden JF. 1995. DNA sequence Tuller T, Waldman YY, Kupiec M, Ruppin E. 2010. Translation efficiency is evolution: the sounds of silence. Philos Trans R Soc Lond B Biol Sci.
determined by both codon bias and folding energy. Proc Natl Acad Sci U S A. 107:3645–3650.
Sharp PM, Emery LR, Zeng K. 2010. Forces that influence the evolution of Tuller T, et al. 2010. An evolutionarily conserved mechanism for controlling codon bias. Philos Trans R Soc Lond B Biol Sci. 365:1203–1212.
the efficiency of protein translation. Cell 141:344–354.
Sharp PM, Li WH. 1987. The codon adaptation index—a measure of dir- Van Rossum G, Drake F. 2001. Python reference manual. VA: PythonLabs ectional synonymous codon usage bias, and its potential applications.
[cited March 2011]. Available at .
Nucleic Acids Res. 15:1281–1295.
Zhou T, Wilke CO. 2011. Reduced stability of mRNA secondary Shen LX, Basilion JP, Stanton VP Jr. 1999. Single-nucleotide polymorphisms structure near the translation-initiation site in dsDNA viruses. BMC can cause different structural folds of mRNA. Proc Natl Acad Sci U S A.
Evol Biol. 11:59.
Sherman DJ, et al. 2009. Genolevures: protein families and synteny among complete hemiascomycetous yeast proteomes and genomes. NucleicAcids Res. 37:D550–D554.
Associate editor: Bill Martin Genome Biol. Evol. 4(10):1046–1053.
Advance Access publication October 3, 2012

Source: http://theshahlab.org/pdfs/papers/Ding_GBE_2012.pdf