Skip Navigation

This Article
Right arrow Abstract Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (82)
Right arrowRequest Permissions
Right arrow Commercial Re-use Guidelines
for Open Access NAR Content
Google Scholar
Right arrow Articles by Moriyama, E. N.
Right arrow Articles by Powell, J. R.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Moriyama, E. N.
Right arrow Articles by Powell, J. R.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

Nucleic Acids Research Pages 3188-3193  

Gene length and codon usage bias in Drosophila melanogaster, Saccharomyces cerevisiae and Escherichia coli
Introduction
Materials And Methods
   Sequence data
   Simulations
Results
   Simulation studies
   Gene length and codon usage bias in D.melanogaster, E.coli and S.cerevisiae
   Gene length and codon usage bias among ribosomal genes
Discussion
Acknowledgements
References

Gene length and codon usage bias in Drosophila melanogaster, Saccharomyces cerevisiae and Escherichia coli

Gene length and codon usage bias in Drosophila melanogaster, Saccharomyces cerevisiae and Escherichia coli

Etsuko N. Moriyama*, Jeffrey R. Powell

Department of Ecology and Evolutionary Biology, Yale University, 165 Prospect Street, New Haven, CT 06520-8106, USA

Received March 17, 1998; Revised and Accepted May 13, 1998

ABSTRACT

The relationship between gene length and synonymous codon usage bias was investigated in Drosophila melanogaster, Escherichia coli and Saccharomyces cerevisiae. Simulation studies indicate that the correlations observed in the three organisms are unlikely to be due to sampling errors or any potential bias in the methods used to measure codon usage bias. The correlation was significantly positive in E.coli genes, whereas negative correlations were obtained for D.melanogaster and S.cerevisiae genes. When only ribosomal protein genes were used, whose expression levels are assumed to be similar, E.coli and S.cerevisiae showed significantly positive correlations. For the two eukaryotes, the distribution of effective number of codons was different in short genes (300-500 bp) compared with longer genes; this was not observed in E.coli. Both positive and negative correlations can be explained by translational selection. Energetically costly longer genes have higher codon usage bias to maximize translational efficiency. Selection may also be acting to reduce the size of highly expressed proteins, and the effect is particularly pronounced in eukaryotes. The different relationships between codon usage bias and gene length observed in prokaryotes and eukaryotes may be the consequence of these different types of selection.

INTRODUCTION

In unicellular organisms (e.g. Escherichia coli and Saccharomyces cerevisiae), synonymous codon usage is biased toward a group of `major' codons. This bias is particularly pronounced in highly expressed genes. It has been found that the codon choice in these organisms is constrained by tRNA availability; `major' or `preferred' codons correspond to the most abundant tRNA species (for a review see 1). Also, an inverse relationship has been found between synonymous substitution rates and codon usage bias in bacterial genes (2,3). These observations are consistent with stronger selection on highly expressed genes to maximize translational efficiency by use of `major' codons (4, but for conflicting results see 5). In lowly expressed genes, on the other hand, variation in mutation bias, probably due to differential DNA repair, seems to play a role in determining codon usage (6).

A negative correlation between codon usage bias and synonymous substitution rates has also been found in Drosophila nuclear genes (7-9). Similar to the case in unicellular organisms, relative tRNA abundance seems to have a relationship with synonymous codon preference in Drosophila genes (10,11). Furthermore, selection acting on Drosophila synonymous codons has been confirmed by population genetics studies (12-14).

Recently, two studies described relationships between codon usage bias and gene length related to translational selection. Eyre-Walker (15) found a positive correlation in E.coli genes expressed at approximately similar levels. He argued that due to a greater cost of producing longer genes, selection in favor of codons which increase translational accuracy should be greater for those genes. On the other hand, Akashi (16) described that when gene lengths were different between homologs in Drosophila melanogaster and Drosophila simulans, the homologs in the former species were longer and less biased than those of the latter species, although the sample size was only six genes and the difference was very small. Relatively reduced selection intensity in D.melanogaster seems to be responsible for such a difference, probably related to a smaller effective population size relative to D.simulans. Energetically costly insertion mutations may have accumulated more in D.melanogaster.

In Moriyama and Powell (10), we examined codon usage bias of D.melanogaster genes. From a sample of >1000 genes, we observed that the average length of the ~10% most highly biased genes was 888 bp, while the ~10% lowest biased genes averaged 2392 bp. This is consistent with observations made by Akashi (16). In this paper, we investigate in more detail the relationship between codon usage bias and gene length in D.melanogaster, comparing it with E.coli and S.cerevisiae.

MATERIALS AND METHODS

Sequence data

Sequence data for D.melanogaster were described in Moriyama and Powell (10). Complete coding sequences 300 bp in length or longer were available for 1001 D.melanogaster genes. The complete sequence of the E.coli genome (strain K-12) is available from GenBank (accession no. U00096), which includes 2463 genes of 300 bp or longer, excluding unidentified ORFs and overlapping sequences. The complete nuclear genome of S.cerevisiae (strain S288C), which is composed of 16 chromosomes, is available by anonymous ftp: genome-ftp.stanford.edu (directory /yeast/genome_seq ). The database includes 2169 identified genes of 300 bp or longer. We used only full-length coding sequences in our analyses.

Simulations

Because sampling errors are affected by sample size (gene length), simulation studies were conducted to examine the behavior of various codon bias measures. We followed the method developed by Wright (17). For a given `true' effective number of codons (ENC), `true' synonymous codon frequencies were estimated for each amino acid. We assumed that one codon is preferred for each amino acid and all other synonymous codons are unpreferred equally. The amino acid composition was held constant; we used the amino acid compositions obtained from D.melanogaster, E.coli and S.cerevisiae genes. For a given `true' ENC, 100 sequences were generated for each of 19 size classes: 70, 100, 150, 180, 240, 300, 360, 410, 440, 500, 560, 620, 660, 800, 1000, 1300, 1600, 2600 and 3600 codons. In addition to ENC, other codon bias measures, scaled [chi]2 ([chi]2/L; 7) and codon adaptation index (CAI; 18), were estimated for each generated sequence. One difference in our simulation from Wright's (17) is that the `observed' ENC was not allowed to be >61. As was done for the actual sequence data, ENC >61 were corrected to 61. While ENC ranges from 61 (no bias) to 20 (the highest bias), CAI and [chi]2/L range from 0 to 1.0 or larger.

All of the statistical analyses were performed using the StatView v.4.51 statistical package (Abacus Concepts Inc.).

RESULTS

Simulation studies

The results of simulation studies are summarized in Table 1. With a given `true' ENC and with a given amino acid composition, 1900 sequences were generated in total. The `true' ENC values, 50 and 41, were chosen based on the observed values from long genes ([ge]4 kb) in D.melanogaster and E.coli. One hundred sequences were generated for each of 19 given lengths (see Materials and Methods). These sequences were grouped arbitrarily into five length categories. Analysis of variance (ANOVA) was conducted among these groups. Correlation coefficients (r and Spearman's r) were calculated between codon usage bias and length of the 1900 sequences. It is clear that [chi]2/L is affected significantly by sequence length (Table 1), as also noted by Wright (17). ENC and CAI are quite insensitive to length. Very similar results were obtained when the amino acid composition of S.cerevisiae genes was used to generate sequences. We also obtained similar results for sequences generated with other ENC values (25, 35, 45, 55 and 60; data not shown).

Table 1. Simulation study for sequence length and codon usage bias
Simulation a Number of
replications		 Mean codon usage bias (SD)
     Length category (kb) ENC CAI [chi]2/L
ENC = 50 (D.melanogaster)
<0.5 300 49.90 (6.78) 0.29 (0.04) 0.30 (0.08)
0.5-1.25 500 50.51 (3.25) 0.29 (0.03) 0.27 (0.06)
1.25-2.0 500 50.41 (2.20) 0.28 (0.02) 0.24 (0.05)
2.0-4.0 300 50.13 (1.67) 0.29 (0.01) 0.24 (0.04)
[ge]4.0 300 50.13 (1.07) 0.28 (0.01) 0.23 (0.03)
Total 1900 50.27 (3.46) 0.29 (0.02) 0.25 (0.06)
Fb   1.96NS 0.89NS 90.36***
rc   -0.02NS(NS) -0.02NS(NS) -0.29***(***)
ENC = 41 (E.coli)
<0.5 300 41.66 (5.85) 0.37 (0.05) 0.46 (0.11)
0.5-1.25 500 41.54 (3.05) 0.37 (0.03) 0.50 (0.09)
1.25-2.0 500 41.40 (1.97) 0.37 (0.02) 0.50 (0.06)
2.0-4.0 300 41.29 (1.53) 0.37 (0.02) 0.50 (0.05)
[ge]4.0 300 41.13 (0.91) 0.37 (0.01) 0.50 (0.03)
Total 1900 41.42 (3.06) 0.37 (0.03) 0.49 (0.08)
Fb   1.49NS 0.97NS 14.46***
rc   -0.05*(NS) -0.03NS(NS) 0.07**(***)
aNucleotide sequences were generated with a given `true' ENC (50 or 41) and a given amino acid composition (D.melanogaster or E.coli).
bF values for ANOVA among length categories.
cCorrelation coefficients for the entire data set. Tests of significance for Spearman's r are shown in parentheses.
*P < 0.05, **P < 0.01, ***P < 0.001, NSnot significant.

Table 2. Simulation study for short sequences and codon usage bias
Length (bp) Mean codon usage bias (SD)
ENC = 50 (D.melanogaster)a ENC = 41 (E.coli)a
ENC CAI ENC CAI
210 46.91 (7.66)[dagger] 0.29 (0.05) 40.08 (7.47) [dagger] 0.37 (0.06)
300 51.29 (6.17) 0.29 (0.04) 41.92 (5.04) 0.37 (0.04)
450 51.49 (5.34) 0.28 (0.03) 43.00 (4.22)[dagger] 0.37 (0.04)
540 50.42 (4.14) 0.29 (0.03) 42.07 (4.19) 0.37 (0.04)
720 50.44 (3.51) 0.29 (0.03) 41.21 (2.19) 0.38 (0.03)
900 50.65 (3.17) 0.28 (0.02) 41.34 (2.69) 0.37 (0.03)
1080 50.50 (2.69) 0.29 (0.02) 41.42 (2.30) 0.37 (0.02)
1230 50.55 (2.56) 0.29 (0.02) 41.66 (2.49) 0.37 (0.03)
Fb 9.04*** 0.32NS 3.79*** 0.70NS
aNucleotide sequences were generated with a given `true' ENC (50 or 41) and a given amino acid composition (D.melanogaster or E.coli). 100 sequences were generated for each length.
bF values for ANOVA among length.
{Mean significantly different from others by Fisher's least significant difference method.
***P < 0.001, NSnot significant.

We further examined the influence of short length on ENC and CAI. In Table 2, each group consists of 100 sequences with a given length. CAI appears to be least affected by short sequence length. ENC fluctuates from giving an overestimate to an underestimate of codon usage bias when sequences are shorter than 500 bp, particularly shorter than 300 bp.


Figure 1. Relationships between codon usage bias (ENC) and gene length (kb) in generated sequences (a), D.melanogaster (b), E.coli (c) and S.cerevisiae (d). The means and standard deviations of ENC are plotted for five length categories. The number of genes included in each category is shown in parentheses. For simulation results (a), open circles are from sequences generated with a `true' ENC = 50 and the amino acid composition of D.melanogaster and closed circles with a `true' ENC = 41 and the amino acid composition of E.coli. Dashed lines indicate the `true' ENC's. For the three organisms (b-d), closed circles and fine lines denote the means and standard deviations of ENC excluding ribosomal protein genes. The numbers of ribosomal protein genes excluded are 25 (b), 35 (c) and 63 (d). Correlation coefficients (r) for the three organisms are given in Table 3. For the simulation studies (a), r = -0.08 (P = 0.001) and -0.09 (P < 0.001) for the `true' ENC = 50 (open circles) and 41 (closed circles) respectively. For both cases, CAI does not have significant correlation (¦r¦ < 0.03, P > 0.2).

Gene length and codon usage bias in D.melanogaster, E.coli and S.cerevisiae

Based on the results of the simulations, we analyzed only genes of 300 bp or longer and used ENC and CAI as codon bias measures. Results for D.melanogaster, E.coli and S.cerevisiae genes are summarized in Figure 1 and Table 3. Figure 2 shows the actual data plotted for the three species. Both ENC and CAI show highly significant heterogeneity and correlations with gene length. Such heterogeneity and correlations do not appear to be due to sampling errors or any potential bias in codon usage measures. Simulation studies show that only when sequences are short (between 300 and 500 bp in length) does ENC tend to underestimate codon bias slightly (Fig. 1a and Table 2). CAI does not show such false heterogeneity and correlations even when sequences are very short. Significant heterogeneity and correlations for both ENC and CAI in the three organisms remains when genes <500 bp are excluded (Table 3). Even when we further exclude genes longer than 4000 bp (sample sizes of such long genes tend to be small, particularly for E.coli) as well as shorter than 500 bp, the results are virtually unchanged (data not shown). The results were also confirmed by a non-parametric Kruskal-Wallis test.

Each of the three organisms examined shows very distinct correlations between codon usage bias and gene length. In accordance with Akashi (16), codon usage bias in D.melanogaster genes correlates negatively with length (represented as a positive correlation with ENC). Highly biased genes in S.cerevisiae are also shorter than the weakly biased genes. On the other hand, E.coli genes show a positive correlation between codon usage bias and length (a negative correlation with ENC), which supports the findings of Eyre-Walker (15). Bacterial genes are generally shorter than eukaryotic genes. The average lengths of coding sequences in our data sets are 1068 bp (ranging up to 4617 bp) for E.coli, 1677 bp (up to 12 279 bp) for S.cerevisiae and 1830 bp (up to 15 444 bp) for D.melanogaster. However, this difference does not seem to contribute to the different correlations seen in these three organisms. As mentioned above, when we used only genes between 500 and 4000 bp in length the patterns of correlation remained similar.

Our data sets include 25 (for D.melanogaster), 35 (for E.coli) and 63 (for S.cerevisiae) ribosomal protein genes. All of them are shorter than 900 bp except for one D.melanogaster gene (1224 bp) and one E.coli gene (1674 bp). Ribosomal protein genes are known to have very high levels of expression and high codon usage bias. The average ENC values of ribosomal protein genes in the three organisms are 35.0 (D.melanogaster), 35.3 (E.coli) and 28.8 (S.cerevisiae). When these ribosomal protein genes are excluded, the patterns of correlation do not change (closed circles in Fig. 1b-d).

We also tried to exclude those which share high sequence similarity (95% or more at the amino acid level) with one or more other genes (presumed duplications), since these genes may have similar length and codon usage bias simply because of their shared evolutionary history. Less than 7% of our data showed such high sequence similarity: 67 D.melanogaster, 97 S.cerevisiae and 28 E.coli genes. Exclusion/inclusion of these sequences did not affect our results (data not shown).

Gene length and codon usage bias among ribosomal genes

Eyre-Walker (15) showed positive correlations between codon usage bias (CAI) and length of E.coli genes by taking advantage of the equimolar presence of ribosomal proteins and also of other multimeric proteins whose gene expression levels are likely similar. We confirmed his results for the 35 ribosomal protein genes in our E.coli data set using both CAI and ENC, although we did not exclude the start and end regions of the genes as he did (in such regions, conflicting selection pressure may be acting, as described later). Codon usage bias of the 63 ribosomal protein genes of S.cerevisiae also has a significant positive correlation with length (r = -0.29, P < 0.02 for ENC; r = 0.34, P < 0.01 for CAI). The correlation in the 25 ribosomal protein genes of D.melanogaster is also positive, though not significant (r = -0.13, P > 0.5 for ENC; r = 0.18, P > 0.3 for CAI). This may be simply due to the relatively small sample size of this data set.

DISCUSSION

The results from ribosomal protein genes are consistent with the translational accuracy model, which is expressed as S [ap] nE, where S is the strength of selection on synonymous codon usage, n is the number of codons and E is the expression level of the protein (15). Synonymous codon usage bias thus depends on the gene length as much as on the expression level. If we assume similar expression levels among these ribosomal protein genes, selective constraints to maximize translational efficiency are stronger in energetically costly longer genes, at least in yeast and E.coli. We should, however, note that some of the eukaryotic ribosomal protein genes have multiple copies and also expression levels may not be the same among them.

The translational accuracy model does not seem to be sufficient to describe the relationships observed in the entire data sets of Drosophila and yeast genes; genes with weaker codon usage bias (these genes are assumed to be weakly constrained) appear to be longer than those with stronger codon usage bias. Selection may also be acting on the size of proteins. Because large proteins are energetically expensive to make, such proteins are disadvantageous if shorter proteins can perform similar functions. Therefore, protein lengths are reduced by selective constraints, while longer proteins are allowed where selective constraints are sufficiently weak. Selective constraints on length seem to be acting differently at the translational and at the transcriptional levels. Drosophila melanogaster has longer coding sequences than D.simulans, whereas there is no difference in length of the introns (16). No consistent difference was found in the length of coding sequences and introns among D.melanogaster, Drosophila virilis and Drosophila pseudoobscura, with one exception, the length of very short introns between D.melanogaster and D.pseudoobscura (19).

Table 3. Relationships between gene length and codon usage bias for three organisms
Statistic Codon usage bias for each organism [no. of genes]
E.coli [2463] S.cerevisiae [2169] D.melanogaster [1001]
ENC CAI ENC CAI ENC CAI
Mean (SD)a 46.25 (5.82) 0.35 (0.11) 48.37 (7.48) 0.22 (0.16) 46.25 (7.00) 0.41 (0.13)
Fb 18.80*** 13.40*** 49.21*** 63.54*** 25.23*** 24.82***
F ([ge]500)c 27.41*** 18.93*** 4.34** 17.15*** 26.83*** 27.87***
rd -0.15*** 0.12*** 0.14*** -0.20*** 0.28*** -0.25***
r ([ge]500)e -0.18*** 0.14*** 0.05* -0.14*** 0.27*** -0.24***
aMean codon usage bias (and SD) for the entire data set. Genes shorter than 300 bp were excluded from the analyses.
bF values for ANOVA among the five length categories. See Table 1 for the categories used.cF values for ANOVA exluding genes shorter than 500 bp.
dCorrelation coefficients for the entire data set.
eCorrelation coefficients excluding genes shorter than 500 bp.
*P < 0.05, **P < 0.01, ***P < 0.001.


Figure 2. Plot of relationship of individual gene length (bp, log scale) and codon usage bias (ENC) for D.melanogaster, S.cerevisiae and E.coli.

A complication arises because under the translational accuracy model, as described above, the strength of selection is proportional to the product of expression level and gene length. If selection is, on the other hand, acting to reduce gene length, how is it possible to see such clear negative correlation in Drosophila and yeast genes between codon usage bias and gene length, but a positive correlation in E.coli genes? We may have to consider two different sources of selection during translation: accuracy and simple length-dependent energy consumption. In the accuracy model, longer genes need to have stronger constraints on codon usage. From the aspect of energy consumption, longer genes are disadvantageous, especially for highly expressed genes. Quantitative data on gene expression levels are not readily available for eukaryotic genes. However, Eyre-Walker (15) compiled data on protein concentration for 46 E.coli genes. Codon usage bias is highly significantly and positively correlated with the protein concentration, as expected (P < 0.0001 for both ENC and CAI). On the other hand, gene length shows a significantly negative correlation with protein concentration (A.Eyre-Walker, personal communication). In E.coli, this correlation does not seem to be sufficient to offset the positive correlation generated by the constraint for translational accuracy.


Figure 3. Distribution of ENC for each length category of E.coli (a), S.cerevisiae (b) and D.melanogaster (c) genes. Only complete genes 300 bp or longer were included.

It is also possible that highly biased genes simply have shorter lengths than weakly biased genes for other functional reasons common to these genes or just by chance. In Figure 3, the distribution of ENC is compared among four length groups (genes with 2000 bp or longer were lumped in one group because of the small sample size for E.coli). For D.melanogaster and S.cerevisiae, the distribution of ENC in short genes appears to be different from that in long genes. Genes between 300 and 500 bp in length have a bimodal (for S.cerevisiae) or nearly flat distribution of ENC. The G test for goodness of fit (with William's correction) showed that ENC of these short genes is homogeneous; genes of each length group were grouped into three subgroups with an equal range of ENC and tested to see if the numbers of genes were significantly heterogeneous among the three subgroups (Gadj = 3.30 and 3.33 for `300-500 bp' genes in D.melanogaster and S.cerevisiae respectively, P > 0.1 for both). However, a difference in codon bias distribution between short and long genes was not observed in E.coli genes.

Although the mechanisms determining synonymous codon usage appear to be similar in E.coli, S.cerevisiae and Drosophila (see Introduction), the pattern of codon usage bias within a gene could be different among them (or possibly between prokaryotic and eukaryotic genomes). In E.coli, codon usage bias is weaker at both the start and end of the genes and conflicting selection pressures, probably due to ribosome binding, have been suggested (20-22). This was attributed to the close proximity of adjacent genes in E.coli. However, only very weak or no intragenic differences in codon usage bias have been detected in yeast genes (22). A recent study for D.melanogaster genes by Kliman and Eyre-Walker (23) showed that G+C% at the third codon position is lower in the start region than the remaining part of the genes. This trend is more pronounced in genes with no introns. Further examination on large data sets may be required. Considering, however, that eukaryotic genes have large intergenic regions and are segmented by introns, it seems that conflicting selection pressure at the start and end regions is not as effective in all eukaryotic genes as in E.coli genes. On the other hand, Akashi (12) found that codon bias varies among synonymous sites when comparing conserved and non-conserved amino acid positions among Drosophila species. Since conserved amino acids showed higher codon bias than non-conserved amino acids, selection for translational accuracy was postulated as a cause of codon bias. Although Hartl et al. (24) found no such evidence in E.coli genes, the positive correlation found between codon usage bias and gene length in E.coli, particularly among the genes expressed approximately equally, appears to be due to selection to avoid missense errors during translation (described in 15). We also found positive correlations when we used only ribosomal protein genes in the yeast and in D.melanogaster, although the correlation was not significant in the latter species.

Therefore, selective constraints on codon usage bias appear to be stronger in energetically costly longer genes to ensure efficient translation by minimizing misincorporation and maximizing translational accuracy. Selection also constrains gene length to reduce the expense of translation. Gene length is likely to be maximally optimized in the compact genome of E.coli, while in larger eukaryotic genomes there appears to be room for such selection to act. What we observed in the three organisms is probably the consequence of these different types of selective constraints. Further detailed analyses on other genomes are evidently necessary. Depending on the intensity of the selection constraints, it may also be possible to see no correlation between gene length and codon usage bias in other eukaryotic genomes. It will be of great interest to see if the contrasted correlations between prokaryotes and eukaryotes can be seen in other organisms.

ACKNOWLEDGEMENTS

We thank P.M.Sharp for his comments on an early version of the manuscript and D.L.Hartl for helpful discussions. Adam Eyre-Walker and anonymous reviewers gave us many critical comments which improved the manuscript significantly. This work was supported by NSF grant DEB 9318836 to J.R.P.

REFERENCES

1. Ikemura,T. (1992) In Hatfield,D.L., Lee,B.J. and Pirtle,R.M. (eds), Transfer RNA in Protein Synthesis. CRC Press, Boca Raton, FL, pp. 87-111.

2. Sharp,P.M. and Li,W.-H. (1987) Mol. Biol. Evol., 4, 222-230. MEDLINE Abstract

3. Sharp,P.M. (1991) J. Mol. Evol., 33, 23-33. MEDLINE Abstract

4. Bulmer,M. (1991) Genetics, 129, 897-907. MEDLINE Abstract

5. Eyre-Walker,A. and Bulmer,M. (1995) Genetics, 140, 1407-1412. MEDLINE Abstract

6. Deschavanne,P. and Filipski,J. (1995) Nucleic Acids Res., 23, 1350-1353. MEDLINE Abstract

7. Shields,D.C., Sharp,P.M., Higgins,D.G. and Wright,F. (1988) Mol. Biol. Evol., 5, 704-716. MEDLINE Abstract

8. Sharp,P.M. and Li,W.-H. (1989) J. Mol. Evol., 28, 398-402. MEDLINE Abstract

9. Powell,J.R. and Moriyama,E.N. (1997) Proc. Natl. Acad. Sci. USA, 94, 7784-7790. MEDLINE Abstract

10. Moriyama, E.N. and Powell, J.R. (1997) J. Mol. Evol., 45, 514-523. MEDLINE Abstract

11. Sharp,P.M. and Lloyd,A.T. (1993) In Maroni,G. (ed.), An Atlas of Drosophila Genes: Sequences and Molecular Features. Oxford University Press, New York, NY, pp. 378-397.

12. Akashi,H. (1994) Genetics, 136, 927-935. MEDLINE Abstract

13. Akashi,H. (1995) Genetics, 139, 1067-1076. MEDLINE Abstract

14. Kliman,R.M. and Hey,J. (1993) Mol. Biol. Evol., 10, 1239-1258. MEDLINE Abstract

15. Eyre-Walker,A. (1996) Mol. Biol. Evol., 13, 864-672. MEDLINE Abstract

16. Akashi,H. (1996) Genetics, 144, 1297-1307. MEDLINE Abstract

17. Wright,F. (1990) Gene, 87, 23-29. MEDLINE Abstract

18. Sharp,P.M. and Li,W.-H. (1987) Nucleic Acids Res., 15, 1281-1295. MEDLINE Abstract

19. Moriyama,E.N., Petrov,D.A. and Hartl,D.L. (1998) Mol. Biol. Evol., 15, 770-773. MEDLINE Abstract

20. Eyre-Walker,A. and Bulmer,M. (1993) Nucleic Acids Res., 21, 4599-4603. MEDLINE Abstract

21. Eyre-Walker,A. (1996) J. Mol. Evol., 42, 73-78. MEDLINE Abstract

22. Bulmer,M. (1988) J. Theor. Biol., 133, 67-71. MEDLINE Abstract

23. Kliman,R.M. and Eyre-Walker,A. (1998) J. Mol. Evol., 46, 534-541. MEDLINE Abstract

24. Hartl,D.L., Moriyama,E.N. and Sawyer,S.A. (1994) Genetics, 138, 227-234. MEDLINE Abstract

*To whom correspondence should be addressed. Tel: +1 203 432 3886; Fax: +1 203 432 6066; Email: moriyama@peaplant.biology.yale.edu

This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 16 Jun 1998
Copyright©Oxford University Press, 1998.
Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 Nucleic Acids ResHome page
Y. Huang, E. V. Koonin, D. J. Lipman, and T. M. Przytycka
Selection for minimization of translational frameshifting errors as a factor in the evolution of codon usage
Nucleic Acids Res., November 1, 2009; 37(20): 6799 - 6810.
[Abstract] [Full Text] [PDF]

Home page
 DNA ResHome page
U. Roymondal, S. Das, and S. Sahoo
Predicting Gene Expression Level from Relative Codon Usage Bias: An Application to Escherichia coli Genome
DNA Res, February 1, 2009; 16(1): 13 - 30.
[Abstract] [Full Text] [PDF]

Home page
 Mol Biol EvolHome page
S. Vicario, C. E. Mason, K. P. White, and J. R. Powell
Developmental Stage and Level of Codon Usage Bias in Drosophila
Mol. Biol. Evol., November 1, 2008; 25(11): 2269 - 2277.
[Abstract] [Full Text] [PDF]

Home page
 J. Virol.Home page
J.-M. Hu, H.-C. Fu, C.-H. Lin, H.-J. Su, and H.-H. Yeh
Reassortment and Concerted Evolution in Banana Bunchy Top Virus Genomes
J. Virol., February 15, 2007; 81(4): 1746 - 1761.
[Abstract] [Full Text] [PDF]

Home page
 Mol Biol EvolHome page
N. Stoletzki and A. Eyre-Walker
Synonymous Codon Usage in Escherichia coli: Selection for Translational Accuracy
Mol. Biol. Evol., February 1, 2007; 24(2): 374 - 381.
[Abstract] [Full Text] [PDF]

Home page
 Mol Biol EvolHome page
A. D. Cutter, J. D. Wasmuth, and M. L. Blaxter
The Evolution of Biased Codon and Amino Acid Usage in Nematode Genomes
Mol. Biol. Evol., December 1, 2006; 23(12): 2303 - 2315.
[Abstract] [Full Text] [PDF]

Home page
 Proc. Natl. Acad. Sci. USAHome page
J. M. Comeron
Weak selection and recent mutational changes influence polymorphic synonymous mutations in humans
PNAS, May 2, 2006; 103(18): 6940 - 6945.
[Abstract] [Full Text] [PDF]

Home page
 Mol Biol EvolHome page
J. M. Comeron and T. B. Guthrie
Intragenic Hill-Robertson Interference Influences Selection Intensity on Synonymous Mutations in Drosophila
Mol. Biol. Evol., December 1, 2005; 22(12): 2519 - 2530.
[Abstract] [Full Text] [PDF]

Home page
 MicrobiologyHome page
M. J. Nicholson, M. K. Theodorou, and J. L. Brookman
Molecular analysis of the anaerobic rumen fungus Orpinomyces - insights into an AT-rich genome
Microbiology, January 1, 2005; 151(1): 121 - 133.
[Abstract] [Full Text] [PDF]

Home page
 Mol Biol EvolHome page
D. Wang, M. Hsieh, and W.-H. Li
A General Tendency for Conservation of Protein Length Across Eukaryotic Kingdoms
Mol. Biol. Evol., January 1, 2005; 22(1): 142 - 147.
[Abstract] [Full Text] [PDF]

Home page
 GeneticsHome page
H. Qin, W. B. Wu, J. M. Comeron, M. Kreitman, and W.-H. Li
Intragenic Spatial Patterns of Codon Usage Bias in Prokaryotic and Eukaryotic Genomes
Genetics, December 1, 2004; 168(4): 2245 - 2260.
[Abstract] [Full Text] [PDF]

Home page
 GeneticsHome page
J. M. Comeron
Selective and Mutational Patterns Associated With Gene Expression in Humans: Influences on Synonymous Composition and Intron Presence
Genetics, July 1, 2004; 167(3): 1293 - 1304.
[Abstract] [Full Text] [PDF]

Home page
 Genome ResHome page
A. D. Cutter, B. A. Payseur, T. Salcedo, A. M. Estes, J. M. Good, E. Wood, T. Hartl, H. Maughan, J. Strempel, B. Wang, et al.
Molecular Correlates of Genes Exhibiting RNAi Phenotypes in Caenorhabditis elegans
Genome Res., December 1, 2003; 13(12): 2651 - 2657.
[Abstract] [Full Text] [PDF]

Home page
 Genome ResHome page
A. O. Urrutia and L. D. Hurst
The Signature of Selection Mediated by Expression on Human Genes
Genome Res., October 1, 2003; 13(10): 2260 - 2264.
[Abstract] [Full Text] [PDF]

Home page
 GeneticsHome page
H. Akashi
Translational Selection and Yeast Proteome Evolution
Genetics, August 1, 2003; 164(4): 1291 - 1303.
[Abstract] [Full Text] [PDF]

Home page
 Mol Biol EvolHome page
G. Marais and G. Piganeau
Hill-Robertson Interference is a Minor Determinant of Variations in Codon Bias Across Drosophila melanogaster and Caenorhabditis elegans Genomes
Mol. Biol. Evol., September 1, 2002; 19(9): 1399 - 1406.
[Abstract] [Full Text] [PDF]

Home page
 Mol Biol EvolHome page
J. A. Novembre
Accounting for Background Nucleotide Composition When Measuring Codon Usage Bias
Mol. Biol. Evol., August 1, 2002; 19(8): 1390 - 1394.
[Full Text] [PDF]

Home page
 GeneticsHome page
J. M. Comeron and M. Kreitman
Population, Evolutionary and Genomic Consequences of Interference Selection
Genetics, May 1, 2002; 161(1): 389 - 410.
[Abstract] [Full Text] [PDF]

Home page
 Nucleic Acids ResHome page
A. Fedorov, S. Saxonov, and W. Gilbert
Regularities of context-dependent codon bias in eukaryotic genes
Nucleic Acids Res., March 1, 2002; 30(5): 1192 - 1197.
[Abstract] [Full Text] [PDF]

Home page
 Mol Biol EvolHome page
C. Pal, B. Papp, and L. D. Hurst
Does the Recombination Rate Affect the Efficiency of Purifying Selection? The Yeast Genome Provides a Partial Answer
Mol. Biol. Evol., December 1, 2001; 18(12): 2323 - 2326.
[Full Text] [PDF]

Home page
 GeneticsHome page
M. A. Jensen, H. L. True, Y. O. Chernoff, and S. Lindquist
Molecular Population Genetics and Evolution of a Prion-like Protein in Saccharomyces cerevisiae
Genetics, October 1, 2001; 159(2): 527 - 535.
[Abstract] [Full Text] [PDF]

Home page
 Nucleic Acids ResHome page
A. Fedorov, S. Saxonov, L. Fedorova, and I. Daizadeh
Comparison of intron-containing and intron-lacking human genes elucidates putative exonic splicing enhancers
Nucleic Acids Res., April 1, 2001; 29(7): 1464 - 1469.
[Abstract] [Full Text] [PDF]

Home page
 GeneticsHome page
G. A. T. McVean and B. Charlesworth
The Effects of Hill-Robertson Interference Between Weakly Selected Mutations on Patterns of Molecular Evolution and Variation
Genetics, June 1, 2000; 155(2): 929 - 944.
[Abstract] [Full Text]

Home page
 Mol Biol EvolHome page
E. Lerat, C. Biemont, and P. Capy
Codon Usage and the Origin of P Elements
Mol. Biol. Evol., March 1, 2000; 17(3): 467 - 468.
[Full Text] [PDF]

Home page
 GeneticsHome page
N. G. C. Smith and L. D. Hurst
The Causes of Synonymous Rate Variation in the Rodent Genome: Can Substitution Rates Be Used to Estimate the Sex Bias in Mutation Rate?
Genetics, June 1, 1999; 152(2): 661 - 673.
[Abstract] [Full Text]

Home page
 GeneticsHome page
A. Llopart and M. Aguadé
Synonymous Rates at the RpII215 Gene of Drosophila: Variation Among Species and Across the Coding Region
Genetics, May 1, 1999; 152(1): 269 - 280.
[Abstract] [Full Text]

Home page
 Proc. Natl. Acad. Sci. USAHome page
L. Duret and D. Mouchiroud
Expression pattern and, surprisingly, gene length shape codon usage in Caenorhabditis, Drosophila, and Arabidopsis
PNAS, April 13, 1999; 96(8): 4482 - 4487.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (82)
Right arrowRequest Permissions
Right arrow Commercial Re-use Guidelines
for Open Access NAR Content
Google Scholar
Right arrow Articles by Moriyama, E. N.
Right arrow Articles by Powell, J. R.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Moriyama, E. N.
Right arrow Articles by Powell, J. R.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?