Random genetic drift sets an upper limit on mRNA splicing accuracy in metazoans

Most eukaryotic genes undergo alternative splicing (AS), but the overall functional significance of this process remains a controversial issue. It has been noticed that the complexity of organisms (assayed by the number of distinct cell types) correlates positively with their genome-wide AS rate. This has been interpreted as evidence that AS plays an important role in adaptive evolution by increasing the functional repertoires of genomes. However, this observation also fits with a totally opposite interpretation: given that 'complex' organisms tend to have small effective population sizes (Ne), they are expected to be more affected by genetic drift, and hence more prone to accumulate deleterious mutations that decrease splicing accuracy. Thus, according to this 'drift barrier' theory, the elevated AS rate in complex organisms might simply result from a higher splicing error rate. To test this hypothesis, we analyzed 3496 transcriptome sequencing samples to quantify AS in 53 metazoan species spanning a wide range of Ne values. Our results show a negative correlation between Ne proxies and the genome-wide AS rates among species, consistent with the drift barrier hypothesis. This pattern is dominated by low abundance isoforms, which represent the vast majority of the splice variant repertoire. We show that these low abundance isoforms are depleted in functional AS events, and most likely correspond to errors. Conversely, the AS rate of abundant isoforms, which are relatively enriched in functional AS events, tends to be lower in more complex species. All these observations are consistent with the hypothesis that variation in AS rates across metazoans reflects the limits set by drift on the capacity of selection to prevent gene expression errors.

Population designations in biomedical research: Limitations and perspectives

In biomedical research, population differences are of central interest. Variations in the frequency and severity of diseases and in treatment effects among human subpopulation groups are common in many medical conditions. Unfortunately, the practices in terms of subpopulation labeling do not exhibit the level of rigor one would expect in biomedical research, especially when studying multifactorial diseases such as cancer or atherosclerosis. The reporting of population differences in clinical research is characterized by large disparities in practices, and fraught with methodological issues and inconsistencies. The actual designations such as "Black" or "Asian" refer to broad and heterogeneous groups, with a great discrepancy among countries. Moreover, the use of obsolete concepts such as "Caucasian" is unfortunate and imprecise. The use of adequate labeling to reflect the scientific hypothesis needs to be promoted. Furthermore, the use of "race/ethnicity" as a unique cause of human heterogeneity may distract from investigating other factors related to a medical condition, particularly if this label is employed as a proxy for cultural habits, diet, or environmental exposure. In addition, the wide range of opinions among researchers does not facilitate the attempts made for resolving this heterogeneity in labeling. "Race," "ethnicity," "ancestry," "geographical origin," and other similar concepts are saturated with meanings. Even if the feasibility of a global consensus on labeling seems difficult, geneticists, sociologists, anthropologists, and ethicists should help develop policies and practices for the biomedical field.

Massive colonization of protein-coding exons by selfish genetic elements in Paramecium germline genomes

Ciliates are unicellular eukaryotes with both a germline genome and a somatic genome in the same cytoplasm. The somatic macronucleus (MAC), responsible for gene expression, is not sexually transmitted but develops from a copy of the germline micronucleus (MIC) at each sexual generation. In the MIC genome of Paramecium tetraurelia, genes are interrupted by tens of thousands of unique intervening sequences called internal eliminated sequences (IESs), which have to be precisely excised during the development of the new MAC to restore functional genes. To understand the evolutionary origin of this peculiar genomic architecture, we sequenced the MIC genomes of 9 Paramecium species (from approximately 100 Mb in Paramecium aurelia species to >1.5 Gb in Paramecium caudatum). We detected several waves of IES gains, both in ancestral and in more recent lineages. While the vast majority of IESs are single copy in present-day genomes, we identified several families of mobile IESs, including nonautonomous elements acquired via horizontal transfer, which generated tens to thousands of new copies. These observations provide the first direct evidence that transposable elements can account for the massive proliferation of IESs in Paramecium. The comparison of IESs of different evolutionary ages indicates that, over time, IESs shorten and diverge rapidly in sequence while they acquire features that allow them to be more efficiently excised. We nevertheless identified rare cases of IESs that are under strong purifying selection across the aurelia clade. The cases examined contain or overlap cellular genes that are inactivated by excision during development, suggesting conserved regulatory mechanisms. Similar to the evolution of introns in eukaryotes, the evolution of Paramecium IESs highlights the major role played by selfish genetic elements in shaping the complexity of genome architecture and gene expression.

A comparative genomics study of nine Paramecium species reveals successful invasion of genes by transposable elements in their germline genomes, showing that the internal eliminated sequences (IESs) followed an evolutionary trajectory remarkably similar to that of spliceosomal introns.

Evolutionary Plasticity of Mating-Type Determination Mechanisms in Paramecium aurelia Sibling Species

The Paramecium aurelia complex, a group of morphologically similar but sexually incompatible sibling species, is a unique example of the evolutionary plasticity of mating-type systems. Each species has two mating types, O (Odd) and E (Even). Although O and E types are homologous in all species, three different modes of determination and inheritance have been described: genetic determination by Mendelian alleles, stochastic developmental determination, and maternally inherited developmental determination. Previous work in three species of the latter kind has revealed the key roles of the E-specific transmembrane protein mtA and its highly specific transcription factor mtB: type O clones are produced by maternally inherited genome rearrangements that inactivate either mtA or mtB during development. Here we show, through transcriptome analyses in five additional species representing the three determination systems, that mtA expression specifies type E in all cases. We further show that the Mendelian system depends on functional and nonfunctional mtA alleles, and identify novel developmental rearrangements in mtA and mtB which now explain all cases of maternally inherited mating-type determination. Epistasis between these genes likely evolved from less specific interactions between paralogs in the P. aurelia common ancestor, after a whole-genome duplication, but the mtB gene was subsequently lost in three P. aurelia species which appear to have returned to an ancestral regulation mechanism. These results suggest a model accounting for evolutionary transitions between determination systems, and highlight the diversity of molecular solutions explored among sibling species to maintain an essential mating-type polymorphism in cell populations.

DILS: Demographic inferences with linked selection by using ABC

We present DILS, a deployable statistical analysis platform for conducting demographic inferences with linked selection from population genomic data using an Approximate Bayesian Computation framework. DILS takes as input single-population or two-population data sets (multilocus fasta sequences) and performs three types of analyses in a hierarchical manner, identifying: (a) the best demographic model to study the importance of gene flow and population size change on the genetic patterns of polymorphism and divergence, (b) the best genomic model to determine whether the effective size Ne and migration rate N, m are heterogeneously distributed along the genome (implying linked selection) and (c) loci in genomic regions most associated with barriers to gene flow. Also available via a Web interface, an objective of DILS is to facilitate collaborative research in speciation genomics. Here, we show the performance and limitations of DILS by using simulations and finally apply the method to published data on a divergence continuum composed by 28 pairs of Mytilus mussel populations/species.

Bedrock radioactivity influences the rate and spectrum of mutation

All organisms on Earth are exposed to low doses of natural radioactivity but some habitats are more radioactive than others. Yet, documenting the influence of natural radioactivity on the evolution of biodiversity is challenging. Here, we addressed whether organisms living in naturally more radioactive habitats accumulate more mutations across generations using 14 species of waterlice living in subterranean habitats with contrasted levels of radioactivity. We found that the mitochondrial and nuclear mutation rates across a waterlouse species’ genome increased on average by 60% and 30%, respectively, when radioactivity increased by a factor of three. We also found a positive correlation between the level of radioactivity and the probability of G to T (and complementary C to A) mutations, a hallmark of oxidative stress. We conclude that even low doses of natural bedrock radioactivity influence the mutation rate possibly through the accumulation of oxidative damage, in particular in the mitochondrial genome.

How consistent is RAD-seq divergence with DNA-barcode based clustering in insects?

Promoted by the barcoding approach, mitochondrial DNA is more than ever used as a molecular marker to identify species boundaries. Yet, it has been repeatedly argued that it may be poorly suited for this purpose, especially in insects where mitochondria are often associated with invasive intracellular bacteria that may promote their introgression. Here, we inform this debate by assessing how divergent nuclear genomes can be when mitochondrial barcodes indicate very high proximity. To this end, we obtained RAD-seq data from 92 barcode-based species-like units (operational taxonomic units [OTUs]) spanning four insect orders. In 100% of the cases, the observed median nuclear divergence was lower than 2%, a value that was recently estimated as one below which nuclear gene flow is not uncommon. These results suggest that although mitochondria may occasionally leak between species, this process is rare enough in insects to make DNA barcoding a reliable tool for clustering specimens into species-like units.

Population genomics supports clonal reproduction and multiple independent gains and losses of parasitic abilities in the most devastating nematode pest

The root-knot nematodes are the most devastating worms to worldwide agriculture with Meloidogyne incognita being the most widely distributed and damaging species. This parasitic and ecological success seems surprising given its supposed obligatory clonal reproduction. Clonal reproduction has been suspected based on cytological observations but, so far, never confirmed by population genomics data. As a species, M. incognita is highly polyphagous with thousands of host plants. However, different M. incognita isolates present distinct and overlapping patterns of host compatibilities. Historically, four "host races" had been defined as a function of ranges of compatible and incompatible plants. In this study, we used population genomics to assess whether (a) reproduction is actually clonal in this species, (b) the host races follow an underlying phylogenetic signal or, rather represent multiple independent transitions, and (c) how genome variations associate with other important biological traits such as the affected crops and geographical distribution. We sequenced the genomes of 11 M. incognita isolates across Brazil that covered the four host races in replicates. By aligning the genomic reads of these isolates to the M. incognita reference genome assembly, we identified point variations. Analysis of linkage disequilibrium and 4-gametes test showed no evidence for recombination, corroborating the clonal reproduction of M. incognita. The few point variations between the isolates showed no significant association with the host races, the geographical origin of the samples, or the crop on which they have been collected. Addition of isolates from other locations around the world confirmed this lack of underlying phylogenetic signal. This suggests multiple gains and losses of parasitic abilities and adaptations to different environments account for the broad host spectrum and wide geographical distribution of M. incognita and thus to its high economic impact. This surprising adaptability without sex poses both evolutionary and agro-economic challenges.

Evolution of replication origins in vertebrate genomes: rapid turnover despite selective constraints

The replication program of vertebrate genomes is driven by the chromosomal distribution and timing of activation of tens of thousands of replication origins. Genome-wide studies have shown the association of origins with promoters and CpG islands, and their enrichment in G-quadruplex motifs (G4). However, the genetic determinants driving their activity remain poorly understood. To gain insight on the constraints operating on origins, we conducted the first evolutionary comparison of origins across vertebrates. We generated a genome-wide map of chicken origins (the first of a bird genome), and performed a comparison with human and mouse maps. The analysis of intra-species polymorphism revealed a strong depletion of genetic diversity at the core of replication initiation loci. This depletion is not linked to the presence of G4 motifs, promoters or CpG islands. In contrast, we show that origins experienced a rapid turnover during vertebrate evolution, since pairwise comparisons of origin maps revealed that <24% of them are conserved among vertebrates. This study unravels the existence of a novel determinant of origins, the precise functional role of which remains to be determined. Despite the importance of replication initiation for the fitness of organisms, the distribution of origins along vertebrate chromosomes is highly flexible.

Codon Usage Bias in Animals: Disentangling the Effects of Natural Selection, Effective Population Size, and GC-Biased Gene Conversion

Selection on codon usage bias is well documented in a number of microorganisms. Whether codon usage is also generally shaped by natural selection in large organisms, despite their relatively small effective population size (Ne), is unclear. In animals, the population genetics of codon usage bias has only been studied in a handful of model organisms so far, and can be affected by confounding, nonadaptive processes such as GC-biased gene conversion and experimental artefacts. Using population transcriptomics data, we analyzed the relationship between codon usage, gene expression, allele frequency distribution, and recombination rate in 30 nonmodel species of animals, each from a different family, covering a wide range of effective population sizes. We disentangled the effects of translational selection and GC-biased gene conversion on codon usage by separately analyzing GC-conservative and GC-changing mutations. We report evidence for effective translational selection on codon usage in large-Ne species of animals, but not in small-Ne ones, in agreement with the nearly neutral theory of molecular evolution. C- and T-ending codons tend to be preferred over synonymous G- and A-ending ones, for reasons that remain to be determined. In contrast, we uncovered a conspicuous effect of GC-biased gene conversion, which is widespread in animals and the main force determining the fate of AT↔GC mutations. Intriguingly, the strength of its effect was uncorrelated with Ne.

GC-biased gene conversion conceals the prediction of the nearly neutral theory in avian genomes

BACKGROUND

The nearly neutral theory of molecular evolution predicts that the efficacy of natural selection increases with the effective population size. This prediction has been verified by independent observations in diverse taxa, which show that life-history traits are strongly correlated with measures of the efficacy of selection, such as the dN/dS ratio. Surprisingly, avian taxa are an exception to this theory because correlations between life-history traits and dN/dS are apparently absent. Here we explore the role of GC-biased gene conversion on estimates of substitution rates as a potential driver of these unexpected observations.

RESULTS

We analyze the relationship between dN/dS estimated from alignments of 47 avian genomes and several proxies for effective population size. To distinguish the impact of GC-biased gene conversion from selection, we use an approach that accounts for non-stationary base composition and estimate dN/dS separately for changes affected or unaffected by GC-biased gene conversion. This analysis shows that the impact of GC-biased gene conversion on substitution rates can explain the lack of correlations between life-history traits and dN/dS. Strong correlations between life-history traits and dN/dS are recovered after accounting for GC-biased gene conversion. The correlations are robust to variation in base composition and genomic location.

CONCLUSIONS

Our study shows that gene sequence evolution across a wide range of avian lineages meets the prediction of the nearly neutral theory, the efficacy of selection increases with effective population size. Moreover, our study illustrates that accounting for GC-biased gene conversion is important to correctly estimate the strength of selection.

The Red Queen model of recombination hot-spot evolution: a theoretical investigation

In humans and many other species, recombination events cluster in narrow and short-lived hot spots distributed across the genome, whose location is determined by the Zn-finger protein PRDM9. To explain these fast evolutionary dynamics, an intra-genomic Red Queen model has been proposed, based on the interplay between two antagonistic forces: biased gene conversion, mediated by double-strand breaks, resulting in hot-spot extinction, followed by positive selection favouring new PRDM9 alleles recognizing new sequence motifs. Thus far, however, this Red Queen model has not been formalized as a quantitative population-genetic model, fully accounting for the intricate interplay between biased gene conversion, mutation, selection, demography and genetic diversity at the PRDM9 locus. Here, we explore the population genetics of the Red Queen model of recombination. A Wright–Fisher simulator was implemented, allowing exploration of the behaviour of the model (mean equilibrium recombination rate, diversity at the PRDM9 locus or turnover rate) as a function of the parameters (effective population size, mutation and erosion rates). In a second step, analytical results based on self-consistent mean-field approximations were derived, reproducing the scaling relations observed in the simulations. Empirical fit of the model to current data from the mouse suggests both a high mutation rate at PRDM9 and strong biased gene conversion on its targets.

This article is part of the themed issue ‘Evolutionary causes and consequences of recombination rate variation in sexual organisms’.

PRDM9 Methyltransferase Activity Is Essential for Meiotic DNA Double-Strand Break Formation at Its Binding Sites

The programmed formation of hundreds of DNA double-strand breaks (DSBs) is essential for proper meiosis and fertility. In mice and humans, the location of these breaks is determined by the meiosis-specific protein PRDM9, through the DNA-binding specificity of its zinc-finger domain. PRDM9 also has methyltransferase activity. Here, we show that this activity is required for H3K4me3 and H3K36me3 deposition and for DSB formation at PRDM9-binding sites. By analyzing mice that express two PRDM9 variants with distinct DNA-binding specificities, we show that each variant generates its own set of H3K4me3 marks independently from the other variant. Altogether, we reveal several basic principles of PRDM9-dependent DSB site determination, in which an excess of sites are designated through PRDM9 binding and subsequent histone methylation, from which a subset is selected for DSB formation.

Unbiased Estimate of Synonymous and Nonsynonymous Substitution Rates with Nonstationary Base Composition

The measurement of synonymous and nonsynonymous substitution rates (dS and dN) is useful for assessing selection operating on protein sequences or for investigating mutational processes affecting genomes. In particular, the ratio dNdS is expected to be a good proxy for ω, the ratio of fixation probabilities of nonsynonymous mutations relative to that of neutral mutations. Standard methods for estimating dN, dS, or ω rely on the assumption that the base composition of sequences is at the equilibrium of the evolutionary process. In many clades, this assumption of stationarity is in fact incorrect, and we show here through simulations and analyses of empirical data that nonstationarity biases the estimate of dN, dS, and ω. We show that the bias in the estimate of ω can be fixed by explicitly taking into consideration nonstationarity in the modeling of codon evolution, in a maximum likelihood framework. Moreover, we propose an exact method for estimating dN and dS on branches, based on stochastic mapping, that can take into account nonstationarity. This method can be directly applied to any kind of codon evolution model, as long as neutrality is clearly parameterized.

The fitness cost of mis-splicing is the main determinant of alternative splicing patterns

BACKGROUND

Most eukaryotic genes are subject to alternative splicing (AS), which may contribute to the production of protein variants or to the regulation of gene expression via nonsense-mediated messenger RNA (mRNA) decay (NMD). However, a fraction of splice variants might correspond to spurious transcripts and the question of the relative proportion of splicing errors to functional splice variants remains highly debated.

RESULTS

We propose a test to quantify the fraction of AS events corresponding to errors. This test is based on the fact that the fitness cost of splicing errors increases with the number of introns in a gene and with expression level. We analyzed the transcriptome of the intron-rich eukaryote Paramecium tetraurelia. We show that in both normal and in NMD-deficient cells, AS rates strongly decrease with increasing expression level and with increasing number of introns. This relationship is observed for AS events that are detectable by NMD as well as for those that are not, which invalidates the hypothesis of a link with the regulation of gene expression. Our results show that in genes with a median expression level, 92-98% of observed splice variants correspond to errors. We observed the same patterns in human transcriptomes and we further show that AS rates correlate with the fitness cost of splicing errors.

CONCLUSIONS

These observations indicate that genes under weaker selective pressure accumulate more maladaptive substitutions and are more prone to splicing errors. Thus, to a large extent, patterns of gene expression variants simply reflect the balance between selection, mutation, and drift.

Recombination, meiotic expression and human codon usage

Synonymous codon usage (SCU) varies widely among human genes. In particular, genes involved in different functional categories display a distinct codon usage, which was interpreted as evidence that SCU is adaptively constrained to optimize translation efficiency in distinct cellular states. We demonstrate here that SCU is not driven by constraints on tRNA abundance, but by large-scale variation in GC-content, caused by meiotic recombination, via the non-adaptive process of GC-biased gene conversion (gBGC). Expression in meiotic cells is associated with a strong decrease in recombination within genes. Differences in SCU among functional categories reflect differences in levels of meiotic transcription, which is linked to variation in recombination and therefore in gBGC. Overall, the gBGC model explains 70% of the variance in SCU among genes. We argue that the strong heterogeneity of SCU induced by gBGC in mammalian genomes precludes any optimization of the tRNA pool to the demand in codon usage.

Less effective selection leads to larger genomes

The evolutionary origin of the striking genome size variations found in eukaryotes remains enigmatic. The effective size of populations, by controlling selection efficacy, is expected to be a key parameter underlying genome size evolution. However, this hypothesis has proved difficult to investigate using empirical data sets. Here, we tested this hypothesis using 22 de novo transcriptomes and low-coverage genomes of asellid isopods, which represent 11 independent habitat shifts from surface water to resource-poor groundwater. We show that these habitat shifts are associated with higher transcriptome-wide [Formula: see text] After ruling out the role of positive selection and pseudogenization, we show that these transcriptome-wide [Formula: see text] increases are the consequence of a reduction in selection efficacy imposed by the smaller effective population size of subterranean species. This reduction is paralleled by an important increase in genome size (25% increase on average), an increase also confirmed in subterranean decapods and mollusks. We also control for an adaptive impact of genome size on life history traits but find no correlation between body size, or growth rate, and genome size. We show instead that the independent increases in genome size measured in subterranean isopods are the direct consequence of increasing invasion rates by repeat elements, which are less efficiently purged out by purifying selection. Contrary to selection efficacy, polymorphism is not correlated to genome size. We propose that recent demographic fluctuations and the difficulty of observing polymorphism variation in polymorphism-poor species can obfuscate the link between effective population size and genome size when polymorphism data are used alone.

In vivo binding of PRDM9 reveals interactions with noncanonical genomic sites

In mouse and human meiosis, DNA double-strand breaks (DSBs) initiate homologous recombination and occur at specific sites called hotspots. The localization of these sites is determined by the sequence-specific DNA binding domain of the PRDM9 histone methyl transferase. Here, we performed an extensive analysis of PRDM9 binding in mouse spermatocytes. Unexpectedly, we identified a noncanonical recruitment of PRDM9 to sites that lack recombination activity and the PRDM9 binding consensus motif. These sites include gene promoters, where PRDM9 is recruited in a DSB-dependent manner. Another subset reveals DSB-independent interactions between PRDM9 and genomic sites, such as the binding sites for the insulator protein CTCF. We propose that these DSB-independent sites result from interactions between hotspot-bound PRDM9 and genomic sequences located on the chromosome axis.

How and how much does RAD-seq bias genetic diversity estimates?

BACKGROUND

RAD-seq is a powerful tool, increasingly used in population genomics. However, earlier studies have raised red flags regarding possible biases associated with this technique. In particular, polymorphism on restriction sites results in preferential sampling of closely related haplotypes, so that RAD data tends to underestimate genetic diversity.

RESULTS

Here we (1) clarify the theoretical basis of this bias, highlighting the potential confounding effects of population structure and selection, (2) confront predictions to real data from in silico digestion of full genomes and (3) provide a proof of concept toward an ABC-based correction of the RAD-seq bias. Under a neutral and panmictic model, we confirm the previously established relationship between the true polymorphism and its RAD-based estimation, showing a more pronounced bias when polymorphism is high. Using more elaborate models, we show that selection, resulting in heterogeneous levels of polymorphism along the genome, exacerbates the bias and leads to a more pronounced underestimation. On the contrary, spatial genetic structure tends to reduce the bias. We confront the neutral and panmictic model to "ideal" empirical data (in silico RAD-sequencing) using full genomes from natural populations of the fruit fly Drosophila melanogaster and the fungus Shizophyllum commune, harbouring respectively moderate and high genetic diversity. In D. melanogaster, predictions fit the model, but the small difference between the true and RAD polymorphism makes this comparison insensitive to deviations from the model. In the highly polymorphic fungus, the model captures a large part of the bias but makes inaccurate predictions. Accordingly, ABC corrections based on this model improve the estimations, albeit with some imprecisions.

CONCLUSION

The RAD-seq underestimation of genetic diversity associated with polymorphism in restriction sites becomes more pronounced when polymorphism is high. In practice, this means that in many systems where polymorphism does not exceed 2 %, the bias is of minor importance in the face of other sources of uncertainty, such as heterogeneous bases composition or technical artefacts. The neutral panmictic model provides a practical mean to correct the bias through ABC, albeit with some imprecisions. More elaborate ABC methods might integrate additional parameters, such as population structure and selection, but their opposite effects could hinder accurate corrections.

No Evidence That Nitrogen Limitation Influences the Elemental Composition of Isopod Transcriptomes and Proteomes

The field of stoichiogenomics aims at understanding the influence of nutrient limitations on the elemental composition of the genome, transcriptome, and proteome. The 20 amino acids and the 4 nt differ in the number of nutrients they contain, such as nitrogen (N). Thus, N limitation shall theoretically select for changes in the composition of proteins or RNAs through preferential use of N-poor amino acids or nucleotides, which will decrease the N-budget of an organism. While these N-saving mechanisms have been evidenced in microorganisms, they remain controversial in multicellular eukaryotes. In this study, we used 13 surface and subterranean isopod species pairs that face strongly contrasted N limitations, either in terms of quantity or quality. We combined in situ nutrient quantification and transcriptome sequencing to test if N limitation selected for N-savings through changes in the expression and composition of the transcriptome and proteome. No evidence of N-savings was found in the total N-budget of transcriptomes or proteomes or in the average protein N-cost. Nevertheless, subterranean species evolving in N-depleted habitats displayed lower N-usage at their third codon positions. To test if this convergent compositional change was driven by natural selection, we developed a method to detect the strand-asymmetric signature that stoichiogenomic selection should leave in the substitution pattern. No such signature was evidenced, indicating that the observed stoichiogenomic-like patterns were attributable to nonadaptive processes. The absence of stoichiogenomic signal despite strong N limitation within a powerful phylogenetic framework casts doubt on the existence of stoichiogenomic mechanisms in metazoans.

Occurrence of a non deleterious gene conversion event in the BRCA1 gene

The duplication in the primate lineage of a portion of the breast and ovarian cancer susceptibility gene BRCA1 has created a BRCA1 pseudogene 45 kb away. Non-allelic homologous recombination (NAHR) between BRCA1 and BRCA1P1 has generated recurrent deleterious germ-line 37-kb deletions encompassing the first two exons of BRCA1, accounting for several breast and ovarian cancer families in various populations. In principle, NAHR intermediates resolution could also lead through a non-crossover configuration to interlocus gene conversion (IGC), but none had been described as yet. Here, we report for the first time an IGC event identified in a breast and ovarian cancer family involving exactly the same segment as that involved in the 37-kb deletions. Close examination of the consequences of this IGC event showed that it does not impact BRCA1 expression. Detailed analysis of the regions of homology between BRCA1 and its pseudogene revealed the specificity of the segment where recombination systematically occurs.

Quantification of GC-biased gene conversion in the human genome

Much evidence indicates that GC-biased gene conversion (gBGC) has a major impact on the evolution of mammalian genomes. However, a detailed quantification of the process is still lacking. The strength of gBGC can be measured from the analysis of derived allele frequency spectra (DAF), but this approach is sensitive to a number of confounding factors. In particular, we show by simulations that the inference is pervasively affected by polymorphism polarization errors and by spatial heterogeneity in gBGC strength. We propose a new general method to quantify gBGC from DAF spectra, incorporating polarization errors, taking spatial heterogeneity into account, and jointly estimating mutation bias. Applying it to human polymorphism data from the 1000 Genomes Project, we show that the strength of gBGC does not differ between hypermutable CpG sites and non-CpG sites, suggesting that in humans gBGC is not caused by the base-excision repair machinery. Genome-wide, the intensity of gBGC is in the nearly neutral area. However, given that recombination occurs primarily within recombination hotspots, 1%–2% of the human genome is subject to strong gBGC. On average, gBGC is stronger in African than in non-African populations, reflecting differences in effective population sizes. However, due to more heterogeneous recombination landscapes, the fraction of the genome affected by strong gBGC is larger in non-African than in African populations. Given that the location of recombination hotspots evolves very rapidly, our analysis predicts that, in the long term, a large fraction of the genome is affected by short episodes of strong gBGC.

Optimization of multiplexed RADseq libraries using low-cost adaptors

Reduced representation genomics approaches, of which RADseq is currently the most popular form, offer the possibility to produce genome wide data from potentially any species, without previous genomic information. The application of RADseq to highly multiplexed libraries (including numerous specimens, and potentially numerous different species) is however limited by technical constraints. First, the cost of synthesis of Illumina adaptors including molecular identifiers (MIDs) becomes excessive when numerous specimens are to be multiplexed. Second, the necessity to empirically adjust the ratio of adaptors to genomic DNA concentration impedes the high throughput application of RADseq to heterogeneous samples, of variable DNA concentration and quality. In an attempt to solve these problems, we propose here some adjustments regarding the adaptor synthesis. First, we show that the common and unique (MID) parts of adaptors can be synthesized separately and subsequently ligated, which drastically reduces the synthesis cost, and thus allows multiplexing hundreds of specimens. Second, we show that self-ligation of adaptors, which makes the adaptor concentration so critical, can be simply prevented by using unphosphorylated adaptors, which significantly improves the ligation and sequencing yield.

GC-Content evolution in bacterial genomes: the biased gene conversion hypothesis expands

The characterization of functional elements in genomes relies on the identification of the footprints of natural selection. In this quest, taking into account neutral evolutionary processes such as mutation and genetic drift is crucial because these forces can generate patterns that may obscure or mimic signatures of selection. In mammals, and probably in many eukaryotes, another such confounding factor called GC-Biased Gene Conversion (gBGC) has been documented. This mechanism generates patterns identical to what is expected under selection for higher GC-content, specifically in highly recombining genomic regions. Recent results have suggested that a mysterious selective force favouring higher GC-content exists in Bacteria but the possibility that it could be gBGC has been excluded. Here, we show that gBGC is probably at work in most if not all bacterial species. First we find a consistent positive relationship between the GC-content of a gene and evidence of intra-genic recombination throughout a broad spectrum of bacterial clades. Second, we show that the evolutionary force responsible for this pattern is acting independently from selection on codon usage, and could potentially interfere with selection in favor of optimal AU-ending codons. A comparison with data from human populations shows that the intensity of gBGC in Bacteria is comparable to what has been reported in mammals. We propose that gBGC is not restricted to sexual Eukaryotes but also widespread among Bacteria and could therefore be an ancestral feature of cellular organisms. We argue that if gBGC occurs in bacteria, it can account for previously unexplained observations, such as the apparent non-equilibrium of base substitution patterns and the heterogeneity of gene composition within bacterial genomes. Because gBGC produces patterns similar to positive selection, it is essential to take this process into account when studying the evolutionary forces at work in bacterial genomes.

Classical population genetics models indicate that the efficiency of selection, and hence adaptation, depends on a number of non-selective factors, such as the size of a population or the intensity of recombination. In the last 10 years, evidence has accumulated that another mechanism called GC-Biased Gene Conversion (gBGC) can interfere with selection and even mimic its effects. This phenomenon, which arises from a particularity of the recombination machinery, was first thought to be restricted to sexual eukaryotic organisms. Here, we show that this mechanism probably exists in Bacteria and has a strong impact on their genome evolution. This discovery not only explains many previously unconnected features of bacterial genome evolution, but also highlights the importance of non-adaptive evolutionary processes in Bacteria.

Estimation of the RNU2 macrosatellite mutation rate by BRCA1 mutation tracing

Large tandem repeat sequences have been poorly investigated as severe technical limitations and their frequent absence from the genome reference hinder their analysis. Extensive allelotyping of this class of variation has not been possible until now and their mutational dynamics are still poorly known. In order to estimate the mutation rate of a macrosatellite, we analysed in detail the RNU2 locus, which displays at least 50 different alleles containing 5-82 copies of a 6.1 kb repeat unit. Mining data from the 1000 Genomes Project allowed us to precisely estimate copy numbers of the RNU2 repeat unit using read depth of coverage. This further revealed significantly different mean values in various recent modern human populations, favoring a scenario of fast evolution of this locus. Its proximity to a disease gene with numerous founder mutations, BRCA1, within the same linkage disequilibrium block, offered the unique opportunity to trace RNU2 arrays over a large timescale. Analysis of the transmission of RNU2 arrays associated with one ‘private’ mutation in an extended kindred and four founder mutations in multiple kindreds gave an estimation by maximum likelihood of 5 × 10⁻³ mutations per generation, which is close to that of microsatellites.

Strong heterogeneity in mutation rate causes misleading hallmarks of natural selection on indel mutations in the human genome

Elucidating the mechanisms of mutation accumulation and fixation is critical to understand the nature of genetic variation and its contribution to genome evolution. Of particular interest is the effect of insertions and deletions (indels) on the evolution of genome landscapes. Recent population-scaled sequencing efforts provide unprecedented data for analyzing the relative impact of selection versus nonadaptive forces operating on indels. Here, we combined McDonald-Kreitman tests with the analysis of derived allele frequency spectra to investigate the dynamics of allele fixation of short (1-50 bp) indels in the human genome. Our analyses revealed apparently higher fixation probabilities for insertions than deletions. However, this fixation bias is not consistent with either selection or biased gene conversion and varies with local mutation rate, being particularly pronounced at indel hotspots. Furthermore, we identified an unprecedented number of loci with evidence for multiple indel events in the primate phylogeny. Even in nonrepetitive sequence contexts (a priori not prone to indel mutations), such loci are 60-fold more frequent than expected according to a model of uniform indel mutation rate. This provides evidence of as yet unidentified cryptic indel hotspots. We propose that indel homoplasy, at known and cryptic hotspots, produces systematic errors in determination of ancestral alleles via parsimony and advise caution interpreting classic selection tests given the strong heterogeneity in indel rates across the genome. These results will have great impact on studies seeking to infer evolutionary forces operating on indels observed in closely related species, because such mutations are traditionally presumed homoplasy-free.

Is RAD-seq suitable for phylogenetic inference? An in silico assessment and optimization

INFERRING PHYLOGENETIC RELATIONSHIPS BETWEEN CLOSELY RELATED TAXA CAN BE HINDERED BY THREE FACTORS: (1) the lack of informative molecular variation at short evolutionary timescale; (2) the lack of established markers in poorly studied taxa; and (3) the potential phylogenetic conflicts among different genomic regions due to incomplete lineage sorting or introgression. In this context, Restriction site Associated DNA sequencing (RAD-seq) seems promising as this technique can generate sequence data from numerous DNA fragments scattered throughout the genome, from a large number of samples, and without preliminary knowledge on the taxa under study. However, divergence beyond the within-species level will necessarily reduce the number of conserved and non-duplicated restriction sites, and therefore the number of loci usable for phylogenetic inference. Here, we assess the suitability of RAD-seq for phylogeny using a simulated experiment on the 12 Drosophila genomes, with divergence times ranging from 5 to 63 million years. These simulations show that RAD-seq allows the recovery of the known Drosophila phylogeny with strong statistical support, even for relatively ancient nodes. Notably, this conclusion is robust to the potentially confounding effects of sequencing errors, heterozygosity, and low coverage. We further show that clustering RAD-seq data using the BLASTN and SiLiX programs significantly improves the recovery of orthologous RAD loci compared with previously proposed approaches, especially for distantly related species. This study therefore validates the view that RAD sequencing is a powerful tool for phylogenetic inference.

GC-biased gene conversion in yeast is specifically associated with crossovers: molecular mechanisms and evolutionary significance

GC-biased gene conversion (gBGC) is a process associated with recombination that favors the transmission of GC alleles over AT alleles during meiosis. gBGC plays a major role in genome evolution in many eukaryotes. However, the molecular mechanisms of gBGC are still unknown. Different steps of the recombination process could potentially cause gBGC: the formation of double-strand breaks (DSBs), the invasion of the homologous or sister chromatid, and the repair of mismatches in heteroduplexes. To investigate these models, we analyzed a genome-wide data set of crossovers (COs) and noncrossovers (NCOs) in Saccharomyces cerevisiae. We demonstrate that the overtransmission of GC alleles is specific to COs and that it occurs among conversion tracts in which all alleles are converted from the same donor haplotype. Thus, gBGC results from a process that leads to long-patch repair. We show that gBGC is associated with longer tracts and that it is driven by the nature (GC or AT) of the alleles located at the extremities of the tract. These observations invalidate the hypotheses that gBGC is due to the base excision repair machinery or to a bias in DSB formation and suggest that in S. cerevisiae, gBGC is caused by the mismatch repair (MMR) system. We propose that the presence of nicks on both DNA strands during CO resolution could be the cause of the bias in MMR activity. Our observations are consistent with the hypothesis that gBGC is a nonadaptive consequence of a selective pressure to limit the mutation rate in mitotic cells.

The origin, evolution, and functional impact of short insertion-deletion variants identified in 179 human genomes

Short insertions and deletions (indels) are the second most abundant form of human genetic variation, but our understanding of their origins and functional effects lags behind that of other types of variants. Using population-scale sequencing, we have identified a high-quality set of 1.6 million indels from 179 individuals representing three diverse human populations. We show that rates of indel mutagenesis are highly heterogeneous, with 43%–48% of indels occurring in 4.03% of the genome, whereas in the remaining 96% their prevalence is 16 times lower than SNPs. Polymerase slippage can explain upwards of three-fourths of all indels, with the remainder being mostly simple deletions in complex sequence. However, insertions do occur and are significantly associated with pseudo-palindromic sequence features compatible with the fork stalling and template switching (FoSTeS) mechanism more commonly associated with large structural variations. We introduce a quantitative model of polymerase slippage, which enables us to identify indel-hypermutagenic protein-coding genes, some of which are associated with recurrent mutations leading to disease. Accounting for mutational rate heterogeneity due to sequence context, we find that indels across functional sequence are generally subject to stronger purifying selection than SNPs. We find that indel length modulates selection strength, and that indels affecting multiple functionally constrained nucleotides undergo stronger purifying selection. We further find that indels are enriched in associations with gene expression and find evidence for a contribution of nonsense-mediated decay. Finally, we show that indels can be integrated in existing genome-wide association studies (GWAS); although we do not find direct evidence that potentially causal protein-coding indels are enriched with associations to known disease-associated SNPs, our findings suggest that the causal variant underlying some of these associations may be indels.

Genome-scale coestimation of species and gene trees

Comparisons of gene trees and species trees are key to understanding major processes of genome evolution such as gene duplication and loss. Because current methods to reconstruct phylogenies fail to model the two-way dependency between gene trees and the species tree, they often misrepresent gene and species histories. We present a new probabilistic model to jointly infer rooted species and gene trees for dozens of genomes and thousands of gene families. We use simulations to show that this method accurately infers the species tree and gene trees, is robust to misspecification of the models of sequence and gene family evolution, and provides a precise historic record of gene duplications and losses throughout genome evolution. We simultaneously reconstruct the history of mammalian species and their genes based on 36 completely sequenced genomes, and use the reconstructed gene trees to infer the gene content and organization of ancestral mammalian genomes. We show that our method yields a more accurate picture of ancestral genomes than the trees available in the authoritative database Ensembl.

Evidence for widespread GC-biased gene conversion in eukaryotes

GC-biased gene conversion (gBGC) is a process that tends to increase the GC content of recombining DNA over evolutionary time and is thought to explain the evolution of GC content in mammals and yeasts. Evidence for gBGC outside these two groups is growing but is still limited. Here, we analyzed 36 completely sequenced genomes representing four of the five major groups in eukaryotes (Unikonts, Excavates, Chromalveolates and Plantae). gBGC was investigated by directly comparing GC content and recombination rates in species where recombination data are available, that is, half of them. To study all species of our dataset, we used chromosome size as a proxy for recombination rate and compared it with GC content. Among the 17 species showing a significant relationship between GC content and chromosome size, 15 are consistent with the predictions of the gBGC model. Importantly, the species showing a pattern consistent with gBGC are found in all the four major groups of eukaryotes studied, which suggests that gBGC may be widespread in eukaryotes.

High-quality sequence clustering guided by network topology and multiple alignment likelihood

MOTIVATION

Proteins can be naturally classified into families of homologous sequences that derive from a common ancestor. The comparison of homologous sequences and the analysis of their phylogenetic relationships provide useful information regarding the function and evolution of genes. One important difficulty of clustering methods is to distinguish highly divergent homologous sequences from sequences that only share partial homology due to evolution by protein domain rearrangements. Existing clustering methods require parameters that have to be set a priori. Given the variability in the evolution pattern among proteins, these parameters cannot be optimal for all gene families.

RESULTS

We propose a strategy that aims at clustering sequences homologous over their entire length, and that takes into account the pattern of substitution specific to each gene family. Sequences are first all compared with each other and clustered into pre-families, based on pairwise similarity criteria, with permissive parameters to optimize sensitivity. Pre-families are then divided into homogeneous clusters, based on the topology of the similarity network. Finally, clusters are progressively merged into families, for which we compute multiple alignments, and we use a model selection technique to find the optimal tradeoff between the number of families and multiple alignment likelihood. To evaluate this method, called HiFiX, we analyzed simulated sequences and manually curated datasets. These tests showed that HiFiX is the only method robust to both sequence divergence and domain rearrangements. HiFiX is fast enough to be used on very large datasets.

AVAILABILITY AND IMPLEMENTATION

The Python software HiFiX is freely available at http://lbbe.univ-lyon1.fr/hifix.

Preventing dangerous nonsense: selection for robustness to transcriptional error in human genes

Nonsense Mediated Decay (NMD) degrades transcripts that contain a premature STOP codon resulting from mistranscription or missplicing. However NMD's surveillance of gene expression varies in efficiency both among and within human genes. Previous work has shown that the intron content of human genes is influenced by missplicing events invisible to NMD. Given the high rate of transcriptional errors in eukaryotes, we hypothesized that natural selection has promoted a dual strategy of “prevention and cure” to alleviate the problem of nonsense transcriptional errors. A prediction of this hypothesis is that NMD's inefficiency should leave a signature of “transcriptional robustness” in human gene sequences that reduces the frequency of nonsense transcriptional errors. For human genes we determined the usage of “fragile” codons, prone to mistranscription into STOP codons, relative to the usage of “robust” codons that do not generate nonsense errors. We observe that single-exon genes have evolved to become robust to mistranscription, because they show a significant tendency to avoid fragile codons relative to robust codons when compared to multi-exon genes. A similar depletion is evident in last exons of multi-exon genes. Histone genes are particularly depleted of fragile codons and thus highly robust to transcriptional errors. Finally, the protein products of single-exon genes show a strong tendency to avoid those amino acids that can only be encoded using fragile codons. Each of these observations can be attributed to NMD deficiency. Thus, in the human genome, wherever the “cure” for nonsense (i.e. NMD) is inefficient, there is increased reliance on the strategy of nonsense “prevention” (i.e. transcriptional robustness). This study shows that human genes are exposed to the deleterious influence of transcriptional errors. Moreover, it suggests that gene expression errors are an underestimated phenomenon, in molecular evolution in general and in selection for genomic robustness in particular.

In biological systems mistakes are made constantly because the cellular machinery is complex and error-prone. Mistakes are made in copying DNA for transmission to offspring (“genetic mutations”) but are much more frequent in the day-to-day task of gene expression. Genetic mutations are heritable and therefore have long been the almost exclusive focus of evolutionary genetics research. In contrast, gene expression errors are not inherited and have tended to be disregarded in evolutionary studies. Here we show how human genes have evolved a mechanism to reduce the occurrence of a specific type of gene expression error—transcriptional errors that create premature STOP codons (so-called “nonsense errors”). Nonsense errors are potentially highly toxic for the cell, so natural selection has evolved a strategy called Nonsense Mediated Decay (NMD) to “cure” such errors. However this cure is inefficient. Here we describe how a preventative strategy of “transcriptional robustness” has evolved to decrease the frequency of nonsense errors. Moreover, these “prevention and cure” strategies are used interchangeably—the most transcriptionally robust genes are those for which NMD is most inefficient. Our work implies that gene expression errors play an important role as supporting actors to genetic mutations in molecular evolution, particularly in the evolution of robustness.

Ultra-fast sequence clustering from similarity networks with SiLiX

Background

The number of gene sequences that are available for comparative genomics approaches is increasing extremely quickly. A current challenge is to be able to handle this huge amount of sequences in order to build families of homologous sequences in a reasonable time.

Results

We present the software package SiLiX that implements a novel method which reconsiders single linkage clustering with a graph theoretical approach. A parallel version of the algorithms is also presented. As a demonstration of the ability of our software, we clustered more than 3 millions sequences from about 2 billion BLAST hits in 7 minutes, with a high clustering quality, both in terms of sensitivity and specificity.

Conclusions

Comparing state-of-the-art software, SiLiX presents the best up-to-date capabilities to face the problem of clustering large collections of sequences. SiLiX is freely available at http://lbbe.univ-lyon1.fr/SiLiX.

Effector diversification within compartments of the Leptosphaeria maculans genome affected by Repeat-Induced Point mutations

Fungi are of primary ecological, biotechnological and economic importance. Many fundamental biological processes that are shared by animals and fungi are studied in fungi due to their experimental tractability. Many fungi are pathogens or mutualists and are model systems to analyse effector genes and their mechanisms of diversification. In this study, we report the genome sequence of the phytopathogenic ascomycete Leptosphaeria maculans and characterize its repertoire of protein effectors. The L. maculans genome has an unusual bipartite structure with alternating distinct guanine and cytosine-equilibrated and adenine and thymine (AT)-rich blocks of homogenous nucleotide composition. The AT-rich blocks comprise one-third of the genome and contain effector genes and families of transposable elements, both of which are affected by repeat-induced point mutation, a fungal-specific genome defence mechanism. This genomic environment for effectors promotes rapid sequence diversification and underpins the evolutionary potential of the fungus to adapt rapidly to novel host-derived constraints.

Meiotic recombination favors the spreading of deleterious mutations in human populations

Although mutations that are detrimental to the fitness of organisms are expected to be rapidly purged from populations by natural selection, some disease-causing mutations are present at high frequencies in human populations. Several nonexclusive hypotheses have been proposed to account for this apparent paradox (high new mutation rate, genetic drift, overdominance, or recent changes in selective pressure). However, the factors ultimately responsible for the presence at high frequency of disease-causing mutations are still contentious. Here we establish the existence of an additional process that contributes to the spreading of deleterious mutations: GC-biased gene conversion (gBGC), a process associated with recombination that tends to favor the transmission of GC-alleles over AT-alleles. We show that the spectrum of amino acid-altering polymorphisms in human populations exhibits the footprints of gBGC. This pattern cannot be explained in terms of selection and is evident with all nonsynonymous mutations, including those predicted to be detrimental to protein structure and function, and those implicated in human genetic disease. We present simulations to illustrate the conditions under which gBGC can extend the persistence time of deleterious mutations in a finite population. These results indicate that gBGC meiotic drive contributes to the spreading of deleterious mutations in human populations.

Ftx is a non-coding RNA which affects Xist expression and chromatin structure within the X-inactivation center region

X chromosome inactivation (XCI) is an essential epigenetic process which involves several non-coding RNAs (ncRNAs), including Xist, the master regulator of X-inactivation initiation. Xist is flanked in its 5' region by a large heterochromatic hotspot, which contains several transcription units including a gene of unknown function, Ftx (five prime to Xist). In this article, we describe the characterization and functional analysis of murine Ftx. We present evidence that Ftx produces a conserved functional long ncRNA, and additionally hosts microRNAs (miR) in its introns. Strikingly, Ftx partially escapes X-inactivation and is upregulated specifically in female ES cells at the onset of X-inactivation, an expression profile which closely follows that of Xist. We generated Ftx null ES cells to address the function of this gene. In these cells, only local changes in chromatin marks are detected within the hotspot, indicating that Ftx is not involved in the global maintenance of the heterochromatic structure of this region. The Ftx mutation, however, results in widespread alteration of transcript levels within the X-inactivation center (Xic) and particularly important decreases in Xist RNA levels, which were correlated with increased DNA methylation at the Xist CpG island. Altogether our results indicate that Ftx is a positive regulator of Xist and lead us to propose that Ftx is a novel ncRNA involved in XCI.

Detecting positive selection within genomes: the problem of biased gene conversion

The identification of loci influenced by positive selection is a major goal of evolutionary genetics. A popular approach is to perform scans of alignments on a genome-wide scale in order to find regions evolving at accelerated rates on a particular branch of a phylogenetic tree. However, positive selection is not the only process that can lead to accelerated evolution. Notably, GC-biased gene conversion (gBGC) is a recombination-associated process that results in the biased fixation of G and C nucleotides. This process can potentially generate bursts of nucleotide substitutions within hotspots of meiotic recombination. Here, we analyse the results of a scan for positive selection on genes on branches across the primate phylogeny. We show that genes identified as targets of positive selection have a significant tendency to exhibit the genomic signature of gBGC. Using a maximum-likelihood framework, we estimate that more than 20 per cent of cases of significantly elevated non-synonymous to synonymous substitution rates ratio (d(N)/d(S)), particularly in shorter branches, could be due to gBGC. We demonstrate that in some cases, gBGC can lead to very high d(N)/d(S) (more than 2). Our results indicate that gBGC significantly affects the evolution of coding sequences in primates, often leading to patterns of evolution that can be mistaken for positive selection.

Gene expression in a paleopolyploid: a transcriptome resource for the ciliate Paramecium tetraurelia

Background

The genome of Paramecium tetraurelia, a unicellular model that belongs to the ciliate phylum, has been shaped by at least 3 successive whole genome duplications (WGD). These dramatic events, which have also been documented in plants, animals and fungi, are resolved over evolutionary time by the loss of one duplicate for the majority of genes. Thanks to a low rate of large scale genome rearrangement in Paramecium, an unprecedented large number of gene duplicates of different ages have been identified, making this organism an outstanding model to investigate the evolutionary consequences of polyploidization. The most recent WGD, with 51% of pre-duplication genes still in 2 copies, provides a snapshot of a phase of rapid gene loss that is not accessible in more ancient polyploids such as yeast.

Results

We designed a custom oligonucleotide microarray platform for P. tetraurelia genome-wide expression profiling and used the platform to measure gene expression during 1) the sexual cycle of autogamy, 2) growth of new cilia in response to deciliation and 3) biogenesis of secretory granules after massive exocytosis. Genes that are differentially expressed during these time course experiments have expression patterns consistent with a very low rate of subfunctionalization (partition of ancestral functions between duplicated genes) in particular since the most recent polyploidization event.

Conclusions

A public transcriptome resource is now available for Paramecium tetraurelia. The resource has been integrated into the ParameciumDB model organism database, providing searchable access to the data. The microarray platform, freely available through NimbleGen Systems, provides a robust, cost-effective approach for genome-wide expression profiling in P. tetraurelia. The expression data support previous studies showing that at short evolutionary times after a whole genome duplication, gene dosage balance constraints and not functional change are the major determinants of gene retention.