General lack of global dosage compensation in ZZ/ZW systems? Broadening the perspective with RNA-seq
© Wolf and Bryk; licensee BioMed Central Ltd. 2011
Received: 17 June 2010
Accepted: 1 February 2011
Published: 1 February 2011
Species with heteromorphic sex chromosomes face the challenge of large-scale imbalance in gene dose. Microarray-based studies in several independent male heterogametic XX/XY systems suggest that dosage compensation mechanisms are in place to mitigate the detrimental effects of gene dose differences. However, recent genomic research on female heterogametic ZZ/ZW systems has generated surprising results. In two bird species and one lepidopteran no evidence for a global dosage compensating mechanism has been found. The recent advent of massively parallel RNA sequencing now opens up the possibility to gauge the generality of this observation with a broader phylogenetic sampling. It further allows assessing the validity of microarray-based inference on dosage compensation with a novel technology.
We here expemplify this approach using massively parallel sequencing on barcoded individuals of a bird species, the European crow (Corvus corone), where previously no genetic resources were available. Testing for Z-linkage with quantitative PCR (qPCR,) we first establish that orthology with distantly related species (chicken, zebra finch) can be used as a good predictor for chromosomal affiliation of a gene. We then use a digital measure of gene expression (RNA-seq) on brain transcriptome and confirm a global lack of dosage compensation on the Z chromosome. RNA-seq estimates of male-to-female (m:f) expression difference on the Z compare well to previous microarray-based estimates in birds and lepidopterans. The data further lends support that an up-regulation of female Z-linked genes conveys partial compensation and suggest a relationship between sex-bias and absolute expression level of a gene. Correlation of sex-biased gene expression on the Z chromosome across all three bird species further suggests that the degree of compensation has been partly conserved across 100 million years of avian evolution.
This work demonstrates that the study of dosage compensation has become amenable to species where previously no genetic resources were available. Massively parallele transcriptome sequencing allows re-assessing the degree of dosage compensation with a novel tool in well-studies species and, in addition, gain valuable insights into the generality of mechanisms across independent taxonomic group for both the XX/XY and ZZ/ZW system.
Sex determining mechanisms have evolved many times independently and range from environmental or social clues to genetic hard-wiring [1, 2]. The two most common forms of chromosomal sex determination are the male heterogametic XX/XY system and the ZZ/ZW system, where females are the heterogametic sex. A widely accepted model posits that in the course of evolution an autosomal chromosome acquired a sex determining locus and was henceforth doomed to degrade by mutational silencing and deletions to a point where it now only harbors a handful of essential genes . This results in a large-scale sexual imbalance in gene dose, which, if not countered, should entail severe detrimental effects [4–6]. Compensating mechanisms are thus expected to be in place that reduce the disparity in sex-linked gene expression and thereby mitigate the effect of sexual imbalance in gene dose. It thus did not come as a surprise that diverse epigenetic dosage compensation mechanisms have been discovered in several XX/XY systems as dissimilar as Drosophila, Mus and Caenorhabditis and much progress has been made to understand their evolutionary basis .
The recent embarquement on genome wide-research on female heterogametic ZZ/ZW systems, however, has generated unexpected results that reinstated much interest in the phenomenon [8–10]: two bird species, chicken and zebra finch, seem to entirely lack a global dosage mechanism [11, 12]. Though varying across life stages and tissues [13, 14], m:f ratios of Z-linked gene expression consistently deviate from unity. Microarray studies as well as quantitative PCR (qPCR) based studies suggest values somewhere in between 1 (full compensation) and 2 (absence of compensation). This indicates either a lower efficiency in dosage compensation, or that the buffering capacity of gene interaction networks alone can compensate for dosage differences . Effective regional dosage compensation has been described around the male hyper-methylated (MHM) locus in chicken , but no such regional patterns are found in zebra finch .
Independent evidence for a lack of wholesale dosage compensation comes from the Lepidopteran Bombyx mori and has prompted the idea that species with ZZ/ZW sex chromosomes are fundamentally different from male heterogametic XX/XY systems in the way they deal with differences in gene dosage . At present, such an inference may, however, be premature, as phylogenetically independent sample size is low (two bird species, one moth). It thus remains an open question, if ZZ/ZW systems in general have evolved similar chromatin-modifying mechanism as in XX/XY systems , or if the general buffering capacity of protein interaction networks described for unbalanced dosage of autosomal genes in partial aneuploids [19, 20] sufficiently counteracts the detrimental effects of dosage differences in the heterogametic sex. Broader taxonomic sampling is needed to understand the diverse mechanisms by which ZZ/ZW species deal with gene dosage imbalances, but also to test for the generality of effective compensation mechanisms in further XX/XY systems .
The rapid advances in sequencing technology over the last few years open up this possibility and allow validating microarray-based inference with a novel methodological tool. Xiong et al. 2010  recently applied RNA-seq data in the context of dosage compensation. Using datasets on well established models such as worm and mouse, they reach the conclusion that apparent compensation of gene dosage differences between the X chromosome and autosomes may simply reflect an artifact of the microarray technique. RNA-seq data suggested no dosage compensation of the active X chromosome. This striking result clearly demonstrates that the addition of a novel technology to the study of dosage compensation will prove valuable and has the potential to even overthrow current belief.
RNA-seq further allows addressing dosage compensation in a much broader set of taxa on a genome-wide scale [reviewed in ]. In contrast to microarrays which are traditionally restricted to models organism with rich genomic resources, massively parallel transcriptome sequencing yields sequence information and concomitantly provides a digital measure of gene expression for basically any species . In principle, this allows inferring gene identity by using orthology relationships to related species with annotated genomes and at the same time determine expression levels.
We here demonstrate the applicability of such an approach for the study of dosage compensation in a non-model bird species, the European Crow (Corvus corone). Using RNA-seq data from massively parallel pyrosequencing we examine patterns of autosomal and Z-linked gene expression in brain and address how RNA-seq based estimators of dosage compensation compare to values obtained from microarray or qPCR experiments. We discuss the findings with special reference to the results from the two other bird species and highlight the potential and the limitations of this approach.
Results and discussion
Copy number variation for 21 putatively Z-linked genes as inferred by orthology with zebra finch.
In birds, movement of complete genes (including introns) between sex chromosomes and autosomes seems to be rare, but it does happen. During more than 100 mya of evolution between chicken and zebra finch approx. 2% (15/745) of 1:1 orthologues have moved between the Z chromosome and autosomes (BioMart ENSEMBL Genes 57). Assuming half the divergence time between crows and zebra finch, one would expect syntenic disruption for ~1% of Z-linked genes. Under this premise, finding one aberrant gene in a set of 21 is not unexpected (p = 0.37, Fisher exact test).
Concluding, orthology relationships between zebra finch and chicken are on the whole a good indicator of chromosomal affiliation of a gene in crows and allow for analyses of gene expression on a chromosome-by-chromosome basis. This opens the door to dosage compensation analyses in this species where an annotated genome is not available. Nonetheless, one should be aware that synteny is not perfect, and Z-linkage may be wrongly inferred for a small number of genes. While this may slightly influence the m:f expression ratio on the Z, it will not compromise the general validity of sex-specific expression analyses.
Compared to gonads and other somatic tissue, brain is less susceptible to sex-specific expression patterns  and should thus be better suited to concentrate on effects of dosage differences between the sexes rather than on sex bias in expression owing to sexual selection as e.g. in gonads (but see  describing the weak relationship between sexual antagonism and sex bias in gene expression). We therefore used brain transcriptome data to examine genome-wide patterns of sex bias in expression levels. From a total of 7564 transcripts for which orthology with chicken or zebra finch could be established, we obtained a digital measure of gene expression and inferred the gene's chromosomal location on the basis of orthologous relationships. Due to the high degree of conserved synteny (see above) the general results are unaffected of whether zebra finch or chicken are used as a reference. To reduce sampling variance for genes with very low sequencing coverage, a subset of 3556 moderately to highly expressed genes was used to explore the patterns of sex biased expression across autosomes and sex chromosomes.
Evidence for incomplete dosage compensation
These results lend support to the hypothesis that the overall pattern of ineffective dosage compensation may indeed be a general pattern in birds. Our RNA-seq based estimates of m:f ratios compare rather well with values obtained from microarray studies in other bird species. Ellegren et al. 2007  report a mean m:f ratio of 0.97-1.04 in autosomes and 1.42 for sex chromosomes in somatic tissue of chicken. Itoh et al.  report mean m:f ratios of 0.99/0.99 in autosomes and 1.40/1.23 in sex chromosomes of chicken and zebra finch, respectively. A microarray study on the Lepidopteran Bombyx mori suggests a mean m:f ratio of 1.41 for sex chromosomes of somatic tissue, which is very similar to the estimates in chicken. However, Itoh et al.  suggested that microarray data may underestimate the degree of sex-bias, as estimates of m:f ratios in half a dozen of Z-linked genes were considerably higher (~2) using qPCR. Our study adds information with a third independent method. The concordance of our RNA-seq based estimates with previous microarray-based data from other bird species is in contrast with findings from Xiong et al. 2010  who report a striking difference between microarray and RNA-seq based estimates of X- to autosomal expression ratios. Being a digital measure of expression RNA-seq is accurate across a wider range of expression levels than in microarrays  and estimation variation seems to be considerably lower . Still, it shall not be concealed that that potential statistical biases exist [e.g. ] and that accuracy depends on proper mapping and coverage, and that GC content and the gene set used (which in turn can be a function of coverage) influence expression ratios. More detailed studies contrasting expression patterns from microarray analyses and RNA-seq data are clearly needed in the context of dosage compensation to assess the relative bias associated with each method.
Regardless of which methodology is applied, the degree of dosage compensation suggested by m:f ratio values is below 1.5 for all avian species and one lepidopteran. This is similar to what has been observed in autosomal aneuploids and artificially induced imbalances of autosomal genes where dosage differences are also partially compensated . This may suggest that intrinsic gene network resilience, auto-regulation through negative feedback and competition for limited regulatory factors  may sufficiently counter the imbalance in gene expression due to gene dosage differences in the heterogametic sex. Studies on the effects of autosomal gene dose, e.g. like those recently conducted in Drosophila[19, 20], are highly important to evaluate the contribution of gene network resilience to mitigate potential dosage effects.
Upregulation of female expression
As seen above, m:f ratios of less than two indicate some degree of compensation either by passive resilience of gene interaction networks or by a specifically evolved mechanism. Lower than expected m:f ratios can either be achieved by up-regulation of female or by down-regulation of male Z-linked genes. The former scenario is more plausible though, as in the course of sex chromosome evolution, dosage imbalance in the heterogametic sex should have increasingly experienced counter-selection.
In males, log2 expression levels of Z-linked genes (Z) were lower in tendency, but not significantly different from those of autosomes (A) (mean ± 95% CI: Zm = 6.14 ± 0.24; Am = 6.22 ±, 0.06; Zm:Am = 0.95, 95% CI: 0.77-1.16)., In females, expression of Z-linked genes was significantly lower than of autosomal genes (mean ± 95% CI: Zf = 5.86 ± 0.24; Af = 6.21 ± 0.06.). A Zf:Af ratio of 0.78 (95% CI: 0.64-0.95) is significantly above the ratio given by the number of chromosomes and suggests up-regulation of Z-linked genes relative to autosomal genes in females.
Up-regulation of female Z-linked expression rather than a down-regulation of Z-linked genes in males has also been suggested for zebra finch (Zm:Am = 0.9-1.1; Zf:Af = 0.65-0.9)  and is theoretically predicted by a verbal model from Birchler et al.  that posits a predominant role of negative trans-acting effects. This model is based on the observation that in most cases trans-acting regulatory genes seem to act as repressors rather than activators and reduce the expression of target genes (e.g. white locus in Drosophila) . Reduced dosage in such a trans-acting repressor gene would consequently lead to increased expression of the target gene. Translated into the context of sex-specific dosage compensation this would mean that trans-acting regulatory genes located on the Z should show on average increase the expression of their target genes in the heterogametic sex. In heterogametic females, where Z dosage is reduced compared to ZZ males, target genes are consequently expected to be up-regulated. If target genes are locally confined to the vicinity of the trans-acting gene on the Z chromosome, an up-regulation of Z specific genes in females rather than a down-regulation in males should indeed be observed. This fits the observation in zebra finch  and is consistent with our data. However, the assumption that target genes are confined to the vicinity of the trans acting gene within the same chromosome is arguably too restrictive. If target genes are similarly present on autosomes one would also expect up-regulation of autosomal target genes in heterogametic females. A male-to-female ratio in autosomal expression (Am:Af) of 1.00 (95% CI: 0.93-1.08) estimated from our data, similarly observed for zebra finch and chicken  does not indicate a large-scale up-regulation of autosomal genes in females.
Degree of compensation and level of expression
Previous studies have reported a relationship between the degree of sex bias and the level of expression, but the patterns are inconsistent between species. In zebra finch, male biased genes (m:f ≫ 1) have higher expression than unbiased genes in males, but lower expression in females [14, 16]. A different pattern has been observed in silkworm where there was no difference between biased and unbiased genes in males, but lower expression for male-biased genes in females . In chicken, Melamed & Arnold  find no difference between male-biased and unbiased genes for females, but higher expression levels for male-biased genes in males .
However, the interpretation of the relationship between levels of gene expression and sex bias is not straightforward. High plasticity in expression levels depending on age and tissue and additional effects of sexual selection can create complex patterns. This - together with differences in statistical power - may be reflected in the discrepancy of observations between studies.
Conservation of m:f ratios across avian lineages
Itoh et al 2010  compared m:f ratios across different tissues in chicken embryos and across different tissues and life-stages in zebra finch. Average correlation coefficients were considerably higher for Z linked genes (0.89 and 0.76 for chicken and zebra finch respectively) than for autosomal genes (0.32 and 0.08 respectively). This pattern may point towards a compensation mechanism that is robust across tissues and developmental stages. A similar conclusion can be drawn from cross-species comparisons. While no correlation between zebra finch and chicken could be found for autosomal genes, significant but weak correlations (mean r = 0.19) were found across tissues and life-stages between the two species.
Correlation of m:f ratios for 41 1:1:1 orthologues genes on the Z chromosome (Z) across zebra finch, chicken and crow and for 1265 autosomal (A) 1:1 orthologues between zebra finch and crow.
Z: p = 0.009
A: p < 0.001
Z: p = 0.001
Z: p = 0.060
The fact that the correlation values for both the Z chromosome and autosomes are relatively high compared to previous studies (see above) may partly be explained by using tissue-and age-matched samples. It may, however, also reflect the conservative, rather highly expressed gene set that was selected by mapping success of crow contigs and 1:1 orthology between chicken and zebra finch. Pairwise comparison between chicken and zebra finch with a larger gene set of 109 genes, indicated a slightly lower correlation (r = 0.34) and could point towards the fascinating possibility that expression levels, gene sequence and dosage compensation may co-evolve.
Overall, cross-specific comparison of sex-biased gene expression on the Z chromosome suggests at least some degree of conservation across more than 100 million years of avian evolution and opens up an interesting avenue for comparative studies once m:f ratios are available for additional species.
Localized effects of dosage compensation
It has been suggested in chicken that dosage compensation in birds may be locally restricted to the evolutionary older stratum of the Z chromosome and that the distal end of the younger stratum harbors a concentration of non-compensated genes . Such a mechanism, however, seems to be absent in zebra finch . Under the assumptions that bird chromosomes are not only syntenic , but have undergone only minor intra-chromosomal rearrangements , we should in principle be able to study localized effects of dosage compensation also in non-model bird species. Extensive studies on linkage in two passerine species suggest that only minor rearrangements have occurred on the Z chromosomes between galliforms (chicken) and passerines (flycatcher: , great reed warbler , blue tit , Siberian Jay ), but two other passerine species where linkage information is available suggest large scale rearrangements on the Z [37, 38]. Accordingly, patterns of m:f ratios differ strongly in crows depending on whether chromosomal gene locations are taken from orthologues in chicken or zebra finch (Figure 1 in Additional file 1). This discloses a clear limitation of the comparative approach using genetic non-model species. Even in avian taxa, where synteny is generally high , it will be difficult to pin down localized effects of dosage compensation unless detailed information on gene order is available. Therefore, species where multi-generational samples of related individuals are available and allow pedigree-based approaches for building detailed linkage maps will be prime candidates for future research. Alternatively, obtaining a draft version of a genome with long continuous scaffolds is possible by massively parallel sequencing alone . This opens up the unprecedented possibility of linking gene expression profiles to the physical, if not exact chromosomal location, so at least large scale scaffold information in basically any species.
The lack of global dosage compensation in all ZZ/ZW systems that have so far been investigated reinstated much interest in studying the relationship between gene dosage and sex biased gene expression. So far, the study of dosage compensation has been limited to model organisms where a wealth of genetic tools is available. The advent of next generation sequencing techniques allows broadening the phylogenetic sample which will be central to our understanding how dosage compensation mechanisms evolve. We here exemplify this approach for a bird species, the European crow (Corvus corone), where previously no genomic resources were available.
In a first step, we established via qPCR that synteny to distantly related species with annotated genomes (zebra finch and chicken) is high. This allowed us placing crow contigs on the correct chromosome on the basis of orthology relationships. We then generated 454 transcriptome data for eight male and female individuals from which we derived a measure of digital gene expression (RNA-seq). Our data confirms the global lack of dosage compensation on the Z chromosome. In contrast to striking differences between microarray- and RNA-seq-based X-to-autosomal-expression values recently observed by Xiong et al. 2010 , our estimates of male-to-female expression difference on the Z that compare well to previous microarray-based estimates in birds and lepidopterae. The data further indicates that female up-regulation of Z-linked gene expression is responsible for partial compensation and establishes a link between the level of expression and sex-bias of a gene.
A potential limitation of our data set was the low read coverage, which introduced additional sampling variance and reduced statistical power. However, as technology progresses, this will be countered with an increased amount of sequencing data. Higher coverage data will also allow separating the effects of individual genes and relate them, for instance, to gene ontology terms. A more fundamental limitation, however, is given by the lack of gene order information that can only be inferred from other species with considerable uncertainty. Gene order information is important to address if localized mechanism of dosage compensation exist as has been suggested for the MHM locus in chicken. This challenge can be met by focusing on species with high quality linkage maps or by de novo assembly of a large number of reads into long scaffolds. The latter option now seems to be within reach even for individual labs . Data on more species will eventually allow comparative analyses on the evolution of sex-biased gene expression. An indication from this study that the degree of compensation is partly conserved for over 100 million years of avian evolution makes this exciting prospect.
The Eurasian crow (Corvus corone) has two distinct colour morphs which are generally referred to as semi-species : the all black carrion crow (Corvus [corone] corone) and the gray-coated and gray-bellied hooded crow (Corvus [corone] cornix). Phylogenies based on mtDNA  and several hundred nuclear markers (JBW Wolf, unpublished data), as well as population genetic studies based on RFLPs , microsatellites  and SNP data  fail to separate the two morphs. As dosage compensation is expected to be a feature shared not only on the species level, but potentially for birds and maybe even ZW systems in general, it is warranted to treat carrion and hooded crows as one species for the purpose of this study. We still conducted separate analysis for the morphs which qualitatively yielded the same results.
Wolf et al. 2010  pyrosequenced barcoded cDNA samples from brain of 11 wild caught sub-adult crows (~2 years old) on a Roche 454 FLX sequencer. The resulting 856 675 raw reads from all individuals were assembled de novo into 19 552 contigs from which 7637 orthologues could be identified with zebra finch and chicken. For each of these genes individual expression levels were estimated as reads per million base pairs per kilo base exon (RPKM). For details of sampling regime, laboratory procedures and data processing see Wolf et al. 2010 . Note that the individuals have been sampled in the wild and differences in age and environmental condition is expected to add variance to the expression data set. All results are therefore conservative and are likely to reflect rather strong effect sizes.
Verification of Z-linkage with qPCR
In crows, inferences on chromosomal locations hinge on the plausible assumption of conserved synteny with other bird species . To verify Z-linkage, we picked a subset of 21 crow genes for which orthology could be established with unique genes in zebra finch (Bio Mart, ENSEMBL 57). For each gene we aligned the coding sequence of zebra finch with all orthologous crow contigs mapping to the gene using CODONCODE ALIGNER VERSION 3.0.1. (CodonCode Corporation) and designed one to a maximum of six exonic primers with predicted product lengths of 50-170 bp using the software PRIMER 3  (Table 1 in Additional file 1).
In order to identify an appropriate control gene to which the signals from the putative Z-genes could be normalized, we compared amplification efficiency and variability of 12 putative autosomal genes selected on the basis of orthology to zebra finch and chicken. Two putative autosomal genes, ENSTGUG00000002932 and ENSTGUG00000013338 that showed similar amplification efficiency in both populations of samples, average m:f ratio of 1 and smallest Ct variability across samples were selected as normalizing genes. Results reported in Table 1 were normalized to the signal of ENSTGUG00000002932, and they were subsequently confirmed with ENSTGUG00000013338 (data not shown).
Amplification efficiency was assayed for 12 primer pairs amplifying 6 putative autosomal control genes and for 11 primer pairs amplifying 5 putative Z-linked genes by amplification of 4-log10 serial dilution of the template DNA (100 ng-100 pg). All samples were run in triplicates. For testing of the putative autosomal genes, we used a mix of template DNA from eight samples; for the putative Z-linked genes, we tested 4 samples, 2 from each population and sex. Primers used in these assays and amplification efficiencies are reported in Additional file 1 (Table 2 and 3, respectively).
qPCR reactions were run in triplicates on a total of eight individuals from both carrion and hooded crow populations (4 males and 4 females) with 5 or 10 ng DNA as a template using the FastSYBR protocol and chemistry (Applied Biosystems) on an ABI 7900HT machine (Applied Biosystems). Data was analyzed with the ΔCt method. The investigator was blind to the sexes of the animals tested.
The qPCR approach enabled us to measure the concentration of the putative Z linked genes relative to a putative autosomal reference gene. Deviations from the predicted mean ratio of 1 for males (ZZ), and 0.5 for females (ZW) are expected due to differences in the amplification efficiency of Z-linked primers and the autosomal control. More important than the absolute values for inference of Z-linkage is the ratio of mean male to mean female concentration: E (ZZconc/ZWconc.) = 2. We used independent t-tests for each primer to assess the statistical significance of the difference.
Dosage compensation analysis
a) Individuals and gene set
In our sample of the same eight barcoded individuals used in the qPCR analysis (two adult specimens from each sex and taxon) we find a total of 7564 transcripts for which orthology with chicken or zebra finch can be established. While the longer reads of the 454 technology are suitable for de novo transcriptome assembly, average per gene coverage for this data set is relatively low (6.3 reads per individual and gene). To reduce sampling variance for estimates of mean sex-specific gene expression we therefore restricted the analyses to genes that were expressed at above mean expression levels in at least two males and females and where log2(m:f ratios) did not exceed ±4. This results in a core set of 3556 moderately to highly expressed genes (11.5 reads per individual and gene). Further reduction of the gene set (i.e. all 1301 genes expressed in all eight individuals) would improve the estimates, but introduces a bias against Z-linked genes, as lowly expressed Z linked genes are often not expressed in females (due to a lack of dosage compensation). In the final set of 3556 genes, no such bias can be observed and - apart from a slight over-representation of genes on chromosome 2, 26 and 27 - the fraction of genes included for autosomes and the Z chromosome follows the expectation from zebra finch (Table 4 in Additional file 1). We therefore consider the 3556 gene set used as a good compromise trading-off equal genomic representation and precision in expression estimates.
b) GC content and RNA-seq
PCR amplification during next generation sequencing library preparation is not equally efficient across all genomic regions and depends on sequencing context. A relationship between GC content and RNA-seq estimates of gene expression is therefore expected and has been reported elsewhere . To estimate the effect of GC content on expression we took GC values from zebra finch as a proxy for the GC content of a crow gene, since contigs are often only gene stubs. This assumption seems warranted as GC content is highly correlated even between chicken and zebra finch (r = 0.88, p > 0.001) that are twice as divergent. We find a significant relationship between GC content and male gene expression (F3212 = 254.5, p < 0.001), but no relationship with female expression (F3212 = 0.40 p > 0.05). This results in a significant relationship between GC content and the m:f ratio of gene expression (F3212 = 755.4 p < 0.001). This discrepancy is surprising and may possibly be explained by slight differences in coverage between males (11 reads per gene) and females (12 reads per gene) and the relative impact of GC bias in regions of higher and lower coverage.
GC content varies strongly between micro- and macro-chromosomes in birds and, unless controlled for, the observed GC bias in expression ratios leads to an underestimation of m:f ratios on large chromosomes (including Z) and overestimation of m:f ratios on small micro-chromosomes. As there is no indication for a biologically real difference in m:f ratios between chromosomes [11, 12], the influence of GC was statistically removed. Analyses without controlling for GC qualitatively yield the same results though the m:f ratios of the relative GC-poor Z chromosome are lower.
c) Statistical analysis
Statistical evidence for decreased overall efficiency in dosage compensation was established as follows. First, we took the mean expression value for each sex and gene and calculated the m:f ratio. In the case of complete dosage compensation and equal overall expression in males and females this ratio is expected to be 1 across all autosomes and the Z chromosome alike. In the absence of any dosage compensation the ratio should be 1 for autosomes and 2 for genes on the Z chromosome. Data were normalized by log2 transformation which centres the expected values around 0 (log2(1) = 0) and 1 (log2(2) = 1). Deviations from the 0 baseline were tested in a linear regression framework. M:f expression ratios were regressed against chromosome identity as the categorical variable. By dropping the intercept an F-test tests the null hypothesis that all means are equal to 0 (no dosage compensation) and confidence intervals (Bonferroni corrected for the number of chromosomes) of the parameter estimates can be used to assess which of the chromosome deviates from 0. To account for samples size differences across chromosomes, the regression was weighted by the number of genes.
The Gaussian kernel density function shown in Figure 1 uses the algorithm implemented in the R density () function. The shape of the curve is robust to modifications of the default parameters even when different smoothing kernels were used.
Cross-species comparison of m:f ratios
To prepare the data for the correlations across avian species (Table 2), we started with the chicken Unigene IDs reported in Tomaszycki et al. 2009  for zebra finch (7299 genes) and obtained 1:1 orthologues between zebra finch and chicken Ensembl gene IDs (BioMart ENSEMBL Genes 59). In cases where multiple Unigene IDs corresponded to single Ensembl gene IDs, we took the average m:f ratios for a given Ensembl IDs and obtained a list of 3042 zebra finch - chicken 1:1 orthologues for which we had m:f ratios in zebra finch. We then assigned chicken m:f ratios from Mank et al.  (590 genes, unpublished) and crow m:f ratios (3556 genes) to this set, arriving at core set of 41 1:1:1 orthologues between chicken, zebra finch and crow.
We would like to thank Nicole Thomsen and Elke Bustorf for excellent assistance in the lab and to several members of the Department of Evolutionary Biology for helpful discussions. Funding was provided by VolkswagenStiftung grant I/83496 (JW) and through the Max-Planck Society (JW, JB). We are grateful to Judith Mank and Youichiro Itoh who provided expression data for chicken and zebra finch.
- Mank JE, Avise JC: Evolutionary Diversity and Turn-Over of Sex Determination in Teleost Fishes. Sex Dev. 2009, 3 (2-3): 60-67. 10.1159/000223071.PubMedView ArticleGoogle Scholar
- Sandra GE, Norma MM: Sexual determination and differentiation in teleost fish. Rev Fish Biol Fish. 2010, 20 (1): 101-121. 10.1007/s11160-009-9123-4.View ArticleGoogle Scholar
- Ayling LJ, Griffin DK: The evolution of sex chromosomes. Cytogenet Genome Res. 2002, 99 (1-4): 125-140. 10.1159/000071584.PubMedView ArticleGoogle Scholar
- Forstmeier W, Ellegren H: Trisomy and triploidy are sources of embryo mortality in the zebra finch. Proc R Soc Lond B. 2010, 277 (1694): 2655-2660. 10.1098/rspb.2010.0394.View ArticleGoogle Scholar
- Birchler JA, Riddle NC, Auger DL, Veitia RA: Dosage balance in gene regulation: biological implications. Trends Genet. 2005, 21: 219-226. 10.1016/j.tig.2005.02.010.PubMedView ArticleGoogle Scholar
- Batada NN, Hurst LD: Evolution of chromosome organization driven by selection for reduced gene expression noise. Nature Genet. 2007, 39 (8): 945-949. 10.1038/ng2071.PubMedView ArticleGoogle Scholar
- Vicoso B, Bachtrog D: Progress and prospects toward our understanding of the evolution of dosage compensation. Chromosome Res. 2009, 17 (5): 585-602. 10.1007/s10577-009-9053-y.PubMed CentralPubMedView ArticleGoogle Scholar
- Arnold AP, Itoh Y, Melamed E: A bird's-eye view of sex chromosome dosage compensation. Ann Rev Genom Hum G. 2008, 9: 109-127. 10.1146/annurev.genom.9.081307.164220.View ArticleGoogle Scholar
- Mank JE: The W, X, Y and Z of sex-chromosome dosage compensation. Trends Genet. 2009, 25 (5): 226-233. 10.1016/j.tig.2009.03.005.PubMed CentralPubMedView ArticleGoogle Scholar
- Naurin S, Hansson B, Bensch S, Hasselquist D: Why does dosage compensation differ between XY and ZW taxa?. Trends Genet. 2010, 26 (1): 15-20. 10.1016/j.tig.2009.11.006.PubMedView ArticleGoogle Scholar
- Ellegren H, Hultin-Rosenberg L, Brunstrom B, Dencker L, Kultima K, Scholz B: Faced with inequality: chicken do not have a general dosage compensation of sex-linked genes. Bmc Biology. 2007, 5: 10.1186/1741-7007-5-40.Google Scholar
- Itoh Y, Melamed E, Yang X, Kampf K, Wang S, Yehya N, Van Nas A, Replogle K, Band M, Clayton D, et al: Dosage compensation is less effective in birds than in mammals. J Biol. 2007, 6 (1): 2-10.1186/jbiol53.PubMed CentralPubMedView ArticleGoogle Scholar
- Mank JE, Ellegren H: All dosage compensation is local: Gene-by-gene regulation of sex-biased expression on the chicken Z chromosome. Heredity. 2009, 102 (3): 312-320. 10.1038/hdy.2008.116.PubMedView ArticleGoogle Scholar
- Itoh Y, Replogle K, Kim Y-H, Wade J, Clayton DF, Arnold AP: Sex bias and dosage compensation in the zebra finch versus chicken genomes: General and specialized patterns among birds. Genome Res. 2010, 20 (4): 512-518. 10.1101/gr.102343.109.PubMed CentralPubMedView ArticleGoogle Scholar
- Birchler JA, Fernandez HR, Kavi HH: Commonalities in compensation. BioEssays. 2006, 28: 565-568. 10.1002/bies.20408.PubMedView ArticleGoogle Scholar
- Melamed E, Arnold AP: Regional differences in dosage compensation on the chicken Z chromosome. Genome Biol. 2007, 8 (9): 10.1186/gb-2007-8-9-r202.
- Zha XF, Xia QY, Duan J, Wang CY, He NJ, Xiang ZH: Dosage analysis of Z chromosome genes using microarray in silkworm, Bombyx mori. Insect Biochem Molec. 2009, 39 (5-6): 315-321. 10.1016/j.ibmb.2008.12.003.View ArticleGoogle Scholar
- McQueen HA, Clinton M: Avian sex chromosomes: dosage compensation matters. Chromosome Res. 2009, 17 (5): 687-697. 10.1007/s10577-009-9056-8.PubMedView ArticleGoogle Scholar
- Zhang Y, Malone JH, Powell SK, Periwal V, Spana E, MacAlpine DM, Oliver B: Expression in Aneuploid Drosophila S2 Cells. PLoS Biol. 2010, 8 (2): 10.1371/journal.pbio.1000320.
- McAnally AA, Yampolsky LY: Widespread Transcriptional Autosomal Dosage Compensation in Drosophila Correlates with Gene Expression Level. Genome Biol Evol. 2010, 2010 (0): 44-52. 10.1093/gbe/evp054.View ArticleGoogle Scholar
- Ruiz MF, Esteban MR, Donoro C, Goday C, Sanchez L: Evolution of dosage compensation in diptera: The gene maleless implements dosage compensation in Drosophila (Brachycera suborder) but its homolog in sciara (Nematocera suborder) appears to play no role in dosage compensation. Genetics. 2000, 156 (4): 1853-1865.PubMed CentralPubMedGoogle Scholar
- Xiong YY, Chen XS, Chen ZD, Wang XZ, Shi SH, Wang XQ, Zhang JZ, He XL: RNA sequencing shows no dosage compensation of the active X-chromosome. Nature Genet. 2010, 42 (12): 1043-U1029. 10.1038/ng.711.PubMedView ArticleGoogle Scholar
- Tautz D, Ellegren H, Weigel D: Next Generation Molecular Ecology. Mol Ecol. 2010, 19 (Suppl. 1): 1-3. 10.1111/j.1365-294X.2009.04489.x.PubMedView ArticleGoogle Scholar
- Wang Z, Gerstein M, Snyder M: RNA-Seq: a revolutionary tool for transcriptomics. Nature Rev Genet. 2009, 10 (1): 57-63. 10.1038/nrg2484.PubMed CentralPubMedView ArticleGoogle Scholar
- Ellegren H: Evolutionary stasis: the stable chromosomes of birds. Trends Ecol Evol. 2010, 25 (5): 283-291. 10.1016/j.tree.2009.12.004.PubMedView ArticleGoogle Scholar
- Innocenti P, Morrow EH: The Sexually Antagonistic Genes of Drosophila melanogaster. PLoS Biol. 2010, 8 (3): e1000335-10.1371/journal.pbio.1000335.PubMed CentralPubMedView ArticleGoogle Scholar
- Mane S, Evans C, Cooper K, Crasta O, Folkerts O, Hutchison S, Harkins T, Thierry-Mieg D, Thierry-Mieg J, Jensen R: Transcriptome sequencing of the Microarray Quality Control (MAQC) RNA reference samples using next generation sequencing. BMC Genomics. 2009, 10 (1): 264-10.1186/1471-2164-10-264.PubMed CentralPubMedView ArticleGoogle Scholar
- Oshlack A, Wakefield M: Transcript length bias in RNA-seq data confounds systems biology. Biol Direct. 2009, 4 (1): 14-10.1186/1745-6150-4-14.PubMed CentralPubMedView ArticleGoogle Scholar
- Altug-Teber O, Bonin M, Walter M, Mau-Holzmann UA, Dufke A, Stappert H, Tekesin I, Heilbronner H, Nieselt K, Riess O: Specific transcriptional changes in human fetuses with autosomal trisomies. Cytogenet Genome Res. 2007, 119 (3-4): 171-184. 10.1159/000112058.PubMedView ArticleGoogle Scholar
- Birchler JA, Bhadra U, Bhadra MP, Auger DL: Dosage-dependent gene regulation in multicellular eukaryotes: implications for dosage compensation, aneuploid syndromes, and quantitative traits. Dev Biol. 2001, 234: 275-288. 10.1006/dbio.2001.0262.PubMedView ArticleGoogle Scholar
- Tomaszycki ML, Peabody C, Replogle K, Clayton DF, Tempelman RJ, Wade J: Sexual differentiation of the zebra finch song system: potential roles for sex chromosome genes. BMC Neurosci. 2009, 10: 10.1186/1471-2202-10-24.Google Scholar
- Griffin DK, Robertson LBW, Tempest HG, Skinner BM: The evolution of the avian genome as revealed by comparative molecular cytogenetics. Cytogenet Genome Res. 2007, 117 (1-4): 64-77. 10.1159/000103166.PubMedView ArticleGoogle Scholar
- Backstrom N, Brandstrom M, Gustafsson L, Qvarnstrom A, Cheng H, Ellegren H: Genetic Mapping in a Natural Population of Collared Flycatchers (Ficedula albicollis): Conserved Synteny but Gene Order Rearrangements on the Avian Z Chromosome. Genetics. 2006, 174: 377-386. 10.1534/genetics.106.058917.PubMed CentralPubMedView ArticleGoogle Scholar
- Dawson DA, Burke T, Hansson B, Pandhal J, Hale MC, Hinten GN, Slate J: A predicted microsatellite map of the passerine genome based on chicken-passerine sequence similarity. Mol Ecol. 2006, 15 (5): 1299-1320. 10.1111/j.1365-294X.2006.02803.x.PubMedView ArticleGoogle Scholar
- Hansson B, Ljungqvist M, Dawson DA, Mueller JC, Olano-Marin J, Ellegren H, Nilsson JA: Avian genome evolution: insights from a linkage map of the blue tit (Cyanistes caeruleus). Heredity. 2010, 104 (1): 67-78. 10.1038/hdy.2009.107.PubMedView ArticleGoogle Scholar
- Jaari S, Li MH, Merila J: A first-generation microsatellite-based genetic linkage map of the Siberian jay (Perisoreus infaustus): insights into avian genome evolution. BMC Genomics. 2009, 10: 1-10.1186/1471-2164-10-1.PubMed CentralPubMedView ArticleGoogle Scholar
- Backstrom N, Forstmeier W, Schielzeth H, Mellenius H, Nam K, Bolund E, Webster MT, Ost T, Schneider M, Kempenaers B, et al: The recombination landscape of the zebra finch Taeniopygia guttata genome. Genome Res. 2010, 20 (4): 485-495. 10.1101/gr.101410.109.PubMed CentralPubMedView ArticleGoogle Scholar
- Nanda I, Schlegelmilch K, Haaf T, Schartl M, Schmid M: Synteny conservation of the Z chromosome in 14 avian species (11 families) supports a role for Z dosage in avian sex determination. Cytogenet Genome Res. 2008, 122 (2): 150-156. 10.1159/000163092.PubMedView ArticleGoogle Scholar
- Li RQ, Fan W, Tian G, Zhu HM, He L, Cai J, Huang QF, Cai QL, Li B, Bai YQ, et al: The sequence and de novo assembly of the giant panda genome. Nature. 2010, 463 (7279): 311-317. 10.1038/nature08696.PubMed CentralPubMedView ArticleGoogle Scholar
- Parkin DT, Collinson M, Helbig AJ, Knox AG, Sangster G: The taxonomic status of Carrion and Hooded Crows. British Birds. 2003, 96: 274-290.Google Scholar
- Haring E, Gamauf A, Kryukov AP: Phylogeographic patterns in widespread corvid birds. Mol Phylogenet Evol. 2007, 45 (3): 840-862.PubMedView ArticleGoogle Scholar
- Kryukov AP, Uphyrkina OV, Chelomina GN: Analysis Of Crow Genomes (Corvidae, Passeriformes) From The Zone Of Overlapping Areas Of Habitation And Hybridization. Genetika. 1992, 28 (6): 136-140.Google Scholar
- Haas F, Pointer MA, Saino N, Brodin A, Mundy NI, Hansson B: An analysis of population genetic differentiation and genotype-phenotype association across the hybrid zone of carrion and hooded crows using microsatellites and MC1R. Mol Ecol. 2009, 18 (2): 294-305. 10.1111/j.1365-294X.2008.04017.x.PubMedView ArticleGoogle Scholar
- Wolf JBW, Bayer T, Haubold B, Schilhabel M, Rosenstiel P, Tautz D: Nucleotide divergence versus gene expression differentiation: comparative transcriptome sequencing in natural isolates from the carrion crow and its hybrid zone with the hooded crow. Mol Ecol. 2010, 19: 162-175. 10.1111/j.1365-294X.2009.04471.x.PubMedView ArticleGoogle Scholar
- Backstrom N, Karaiskou N, Leder EH, Gustafsson L, Primmer CR, Qvarnstrom A, Ellegren H: A gene-based genetic linkage map of the collared flycatcher (Ficedula albicollis) reveals extensive synteny and gene-order conservation during 100 million years of avian evolution. Genetics. 2008, 179 (3): 1479-1495. 10.1534/genetics.108.088195.PubMed CentralPubMedView ArticleGoogle Scholar
- Rozen S, Skaletsky HJ: Primer3 on the WWW for general users and for biologist programmers. Bioinformatics Methods and Protocols: Methods in Molecular Biology. Edited by: Krawetz S, Misener S. 2000, Totowa, NJ: Humana Press, 365-386.Google Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.