• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of geneticsGeneticsCurrent IssueInformation for AuthorsEditorial BoardSubscribeSubmit a Manuscript
Genetics. Aug 2008; 179(4): 1903–1917.
PMCID: PMC2516068

Pervasive Sex-Linked Effects on Transcription Regulation As Revealed by Expression Quantitative Trait Loci Mapping in Lake Whitefish Species Pairs (Coregonus sp., Salmonidae)

Abstract

Mapping of expression quantitative trait loci (eQTL) is a powerful means for elucidating the genetic architecture of gene regulation. Yet, eQTL mapping has not been applied toward investigating the regulation architecture of genes involved in the process of population divergence, ultimately leading to speciation events. Here, we conducted an eQTL mapping experiment to compare the genetic architecture of transcript regulation in adaptive traits, differentiating the recently evolved limnetic (dwarf) and benthic (normal) species pairs of lake whitefish. The eQTL were mapped in three data sets derived from an F1 hybrid-dwarf backcrossed family: the entire set of 66 genotyped individuals and the two sexes treated separately. We identified strikingly more eQTL in the female data set (174), compared to both male (54) and combined (33) data sets. The majority of these genes were not differentially expressed between male and female progeny of the backcross family, thus providing evidence for a strong pleiotropic sex-linked effect in transcriptomic regulation. The subtelomeric region of a linkage group segregating in females encompassed >50% of all eQTL, which exhibited the most pronounced additive effects. We also conducted a direct comparison of transcriptomic profiles between pure dwarf and normal progeny reared in controlled conditions. We detected 34 differentially expressed transcripts associated with eQTL segregating only in sex-specific data sets and mostly belonging to functional groups that differentiate dwarf and normal whitefish in natural populations. Therefore, these eQTL are not related to interindividual variation, but instead to the adaptive and historical genetic divergence between dwarf and normal whitefish. This study exemplifies how the integration of genetic and transcriptomic data offers a strong means for dissecting the functional genomic response to selection by separating mapping family-specific effects from genetic factors under selection, potentially involved in the phenotypic divergence of natural populations.

INVESTIGATIONS of quantitative variation in gene transcription have begun to open up the “black box” connecting genotype to phenotype (Gibson and Weir 2005; Prud'homme et al. 2007). Expression quantitative trait loci (eQTL) mapping, which combines transcriptional profiling with linkage mapping methods, allows the investigation of the genetic architecture of gene regulation and quantifies the heritability of these regulation factors (Wayne and McIntyre 2002; Schadt et al. 2003; Petretto et al. 2006; Vuylsteke et al. 2006). This is also a crucial issue in evolutionary biology, as many authors argued that changes in key regulatory genes may have a greater and faster effect than changes in structural genes on early phenotypic diversification (Jacob and Monod 1961; King and Wilson 1975; Purugganan 1998), although this is still a contentious issue (Hoekstra and Coyne 2007). Transcription QTL mapping has almost exclusively been applied to model organisms, including yeast (Brem et al. 2002), human (Morley et al. 2004), mouse, maize (Schadt et al. 2003), and Drosophila (Jin et al. 2002); but see Kirst et al. (2005) for work on the eucalyptus tree. To date, eQTL mapping has not been specifically applied toward investigating the regulation architecture of genes involved in the process of recent population divergence in the wild. However, this should be a fundamental research goal since variation in gene expression within and between natural populations has been recently reported to underlie adaptation (Gibson 2002; Oleksiak et al. 2002, 2005; Bochdanovits et al. 2003; Whitehead and Crawford 2005; Derome and Bernatchez 2006; Derome et al. 2006; Roberge et al. 2007). Investigating the genetic architecture of transcriptome regulation in natural populations or closely related species exhibiting phenotypic divergence related to distinct environmental responses provides a powerful means to reach a more complete understanding of the molecular mechanisms related to the process of adaptive phenotypic divergence, which in turn are involved in the course of speciation events.

Clearly, adaptive trait variation is multifaceted (Reeves and Sherman 1993). In many species, differences between the sexes add another important level of phenotypic variation, as illustrated by substantial sex-linked effects in quantitative studies (Mckay 2001) and, more recently, studies of sex-biased gene transcription (Rantz et al. 2003; McIntyre et al. 2006; Yang et al. 2006; Itoh et al. 2007; Wayne et al. 2007). Indeed, speciation is intimately associated with the evolution of traits related to sex and reproduction, including those affecting hybrid incompatibilities (postzygotic isolation). Also, genes controlling such traits are particularly abundant on the sex chromosomes (Wang et al. 2001; Ellegren and Parsh 2007). However, to our knowledge, little attention has been paid to investigating the genomic basis of sex differences in the context of an adaptive radiation (Pröschel et al. 2006) or the evolutionary consequences of the sex linkage of genes involved in speciation (Servedio and Saetre 2006). However, sex-biased genes have been observed to be among the most rapidly evolving genes in various taxonomic groups such as nematodes, insects, fish, birds, and mammals (reviewed by Singh and Kulathinal 2000; Ellegren and Fridolfsson 2003; Zang et al. 2004; Richards et al. 2005; Ellegren and Parsh 2007). This phenomenon was observed as well for somatic tissues (Parisi et al. 2004; Marinotti et al. 2006; Mank et al. 2007), including muscle (Yang et al. 2006). Furthermore, these sex-biased genes have been reported to cause genetic incompatibilities by disrupting local adaptations or coadaptations within the genome (Parker and Partridge 1998; Fitzpatrick 2004). It thus appears crucial to investigate the genomic basis of sex-specific transcriptional regulation by providing new insights into the role of sex-linked regulation effects differentiating two recently diverged genomes.

Sympatric dwarf (limnetic) and normal (benthic) species pairs of the lake whitefish (Coregonus sp.) constitute a case of recent adaptive radiation that is particularly well suited for investigating the genomic architecture of gene transcription. Geographic isolation during the Pleistocene (500,000–18,000 YBP) led to genetic divergence between whitefish populations inhabiting distinct glacial refuges (Bernatchez and Dodson 1990; Lu et al. 2001). Secondary contact between these evolutionary lineages subsequently occurred ~15,000 YBP within at least six lakes of the Saint John River basin in Maine and southeastern Quebec. Ecological opportunity and character displacement have both contributed to the rapid evolution of a limnetic dwarf species, which has diverged in sympatry from the ancestral benthic normal species (Bernatchez 2004; Landry et al. 2007). Adaptive trait differences observed in the dwarf and normal whitefish dichotomy are supported by several genetically based phenotype–environment associations, including for life history (age of maturity, life span), morphological (size, weight), behavioral (swimming activity), physiological (metabolic rate and growth), and gene expression traits (Lu and Bernatchez 1999; Trudel et al. 2001; Rogers et al. 2002; Rogers and Bernatchez 2005; Derome et al. 2006). Indeed, dwarf exhibits a higher metabolic rate, partly associated with the cost of a more active swimming activity (higher position in water column, direction changes, burst swims), and lower bioenergetic conversion efficiency (growth rate/consumption rate ratio), associated with slower growth and younger age at sexual maturity in dwarf whitefish (Trudel et al. 2001) when compared to normal whitefish (Rogers et al. 2002; Rogers and Bernatchez 2007). Moreover, linkage mapping has been used to document the number and effects of quantitative trait loci (QTL) involved in controlling the expression of these adaptive traits (Rogers et al. 2007), and genome scans performed in natural populations provided evidence that directional selection is maintaining genetic divergence between sympatric dwarf and normal whitefish by restricting gene flow at more than half of these QTL (Rogers and Bernatchez 2007). Finally, functional genomic studies performed in these same natural populations on the white muscle tissue showed that dwarf whitefish overexpressed genes associated with muscle contraction speed, which is consistent with the higher swimming activity of limnetic foraging species (Derome et al. 2006). Thus, the accumulation of genetic differences during the allopatric phase of geographic isolation in conjunction with ecological divergence that subsequently occurred in sympatry led to reproductive isolation between dwarf and normal whitefish species pairs (Lu and Bernatchez 1998; Rogers and Bernatchez 2006).

Here, we investigate the genetic architecture of transcript regulation in the white muscle tissue differentiating the recently evolved dwarf and normal species pairs of lake whitefish by conducting an eQTL mapping study using a 16,006-gene cDNA salmonids-specific microarray (Von Schalburg et al. 2005). Transcript abundance was measured in the same progeny of an F1 hybrid-dwarf backcrossed family for which a genetic map was recently built (Rogers et al. 2007). In this way, both genotypic and transcriptomic information were integrated to identify genes potentially underlying the divergence in adaptive traits between dwarf and normal whitefish, namely in terms of their energy metabolism and swimming activity (Derome et al. 2006). As there is a trade-off in energy allocated to growth vs. swimming activity (Trudel et al. 2001; Rogers and Bernatchez 2007), we paid particular attention to the relationship between eQTL associated with muscle contraction, energetic metabolism, and protein synthesis. Namely, we predicted opposite additive effects between eQTL of transcripts related to growth and swimming activity.

Moreover, to test whether the architecture of transcriptomic profiles segregated between sexes, we compared sex-specific data sets to a combined data set termed “combined,” including all individuals, but excluding the sex effect on transcript levels (see materials and methods). Sexual dimorphism in gene expression may be substantial, as reported recently both for Drosophila (Jin et al. 2002) and for mouse (Yang et al. 2006). Therefore, to distinguish between sex-specific effects (independent from the segregating genetic background) and pleiotropic sex-linked effects potentially involved in the phenotypic divergence of dwarf and normal whitefish (dependent on the segregating genetic background), we contrasted transcripts that exhibited eQTL from expression profile mapping experiments to genes differentially expressed between sexes of the same backcross progeny. In this way, transcripts that exhibited eQTL in a male or a female data set, but were not differentially expressed between progeny of different sexes, were considered as pleiotropic sex-linked effects potentially involved in the adaptive divergence between whitefish ecotypes.

Finally, to further assess whether the eQTL detected were potentially relevant in the context of adaptive divergence, we conducted transcriptomic comparisons of pure dwarf and normal progeny issued from the grandparents of the backcross family and reared in controlled conditions. To identify potential candidate genes for which transcription levels may have evolved under directional selection between dwarf and normal whitefish, we paid particular attention to (i) transcripts associated with eQTL that were differentially expressed between pure dwarf and normal whitefish and (ii) eQTL belonging to functional groups that exhibited significant directional additive effects. The predominance of additive effects (i.e., amount of a given transcript modulated by segregation of alleles between dwarf and normal genomes) in one direction for the loci underlying the quantitative variation of a given trait would provide further evidence for the role of selection in shaping genome architecture (Orr 1998).

MATERIALS AND METHODS

Linkage maps:

The eQTL were mapped onto a hybrid (normal female × dwarf male) × dwarf backcross (BC1) linkage map built by Rogers et al. (2007). Briefly, sex-specific markers segregating in the backcross families were assigned to a linkage group under a LOD of 4 while orders and recombination distances of these groups were determined under the Kosambi mapping function independently for both families using MAPMAKER/EXP (Lander et al. 1987). Defining sex-specific linkage groups was necessary since salmonids exhibit a sex-specific recombination rate due to tetravalent formations in the males during meioses (Sakamoto et al. 2000; Gharbi et al. 2006). The hybrid × dwarf map was built using a total of 389 AFLPs and 23 microsatellites covering a total of 48 linkage groups, including 34 female linkage groups (average = 8.0 loci per linkage group, average distance between markers = 17.3 cM, and a total map length of 2800 cM) and 14 dwarf male linkage groups (average = 8.7 loci per linkage group, average distance between markers = 17.96 cM, and a total map length of 2127.5 cM). The homology of these mapped loci was verified in a second backcross family where 34 linkage groups of the 40n expected showed 83% colinearity among linked loci between both families (Rogers et al. 2007).

Tissue sampling, RNA preparation, labeling, and hybridization:

Dissected white muscle tissue (250–350 mg) was sampled for 66 individuals from the hybrid × dwarf backcross mapping family and stored at −80° until RNA extraction. RNA was extracted according to the Trizol Reagent protocol (GIBCO BRL, Gaithersburg, MD) and quantified with a GeneQuant spectrometer (Pharmacia, Piscataway, NJ), and RNA integrity was verified with a 2100 Bioanalyzer (Agilent). Reverse transcriptase PCR was performed using 15 μg of total RNA per sample following the SuperScript II Reverse Transcriptase protocol (Invitrogen Life Technologies, San Diego). Indirect labeling was performed on individual cDNA following the Array 50 kit protocol (Genisphere). Transcriptome profiles were obtained by using the 16,006-cDNA gene microarray (version 2) developed for Atlantic salmon (Salmo salar) by cGRASP (Von Schalburg et al. 2005) (Consortium for Genomic Research on All Salmonids Project) and successfully tested and applied to other salmonid species, including Coregonus clupeaformis (Derome and Bernatchez 2006; Derome et al. 2006; Rise et al. 2007). Gene identification with the corresponding EST sequence can be found at http://web.uvic.ca/cbr/grasp/. Cross-hybridization is known to be potentially a problem for spotted cDNA microarrays because of sequence polymorphisms between strains or paralogous genes that affect the signal for certain genes. However, in this study, we did not compare Salmo to Coregonus transcripts, but a backcross family of whitefish species pairs that diverged <12,000 YBP. Therefore, differences in cross-hybridization between dwarf and normal whitefish alleles would be negligible and unlikely to affect the results obtained. The transcript levels inferred from fluorescent labels were quantified by scanning microarrays using a ScanArray Express scanner (Packard Bioscience). Spot location and quantification was done with the QuantArray (Perkin-Elmer, Norwalk, CT) software, retaining the mean intensity value for each spot. Aberrant spot signals were removed before analysis and their values were estimated using the “K-Nearest Neighbors Imputer” function implemented in SAM software (Tusher et al. 2001). Genes with intensity less than the mean of the empty spot controls plus 2.5 times their standard deviation were removed from the analysis. According to this criterion, 2255 among the 16,006 transcripts passed this threshold and were considered for subsequent analyses. Raw and processed data sets are available at the GEO website (http://www.ncbi.nlm.nih.gov/geo/).

Experimental design and statistical analyses applied to eQTL experiments:

We used a loop design (Churchill 2002; Yang and Speed 2002) to maximize the number of sampled meioses. Each of 66 samples was technically replicated on two distinct slides, while performing dye swapping (Cy3 and Alexa 647) to estimate the dye intensity variation bias. The loop design allows for the evaluation of twice as many recombined genotypes for a given number of slides as does reference design (Kirst et al. 2005). After correcting for local background, raw intensity values were both log2 transformed and normalized using the regional LOWESS method implemented in the R/MANOVA software (Kerr et al. 2000), available at http://www.jax.org/staff/churchill/labsite/software/Rmaanova. Then, the gene–genotype interaction effect (G) was estimated using the R/MAANOVA software under a mixed-effects model of variance analysis (Cui and Churchill 2003) with the array term kept as a random effect, while genotype, sex, and dye term effects were fixed. Note that the genotype term refers to the individual, as each offspring of the backcross family constitutes a unique event of recombination between the parental genotypes. Two ANOVA models were used, the first one being applied to the whole data set, hereafter termed a combined data set [Yijkgh = μ + Ai + Dj + Tk + Gg +Sh + (AD)ij + (GA)gi + (GD)gj + (GT)gk + (GS)gh + epsilonijkgh], and the second one used for separate sex data sets [Yijkg = μ + Ai + Dj + Tk + Gg + (AD)ij + (GA)gi + (GD)gj + (GT)gk + epsilonijkg] (A, array; D, dye; G, gene; T, genotype; and S, sex; terms in parentheses are interaction terms).

Detections of eQTL:

For each transcript, the estimates of the variation associated with gene–genotype interaction for the 66 individual progeny were used as the phenotypes along with the linkage map to perform genomewide eQTL detection from the combined and sex-specific data sets, using the UNIX version of QTL Cartographer (Batsen et al. 2002).

The combined data set comprised all of the 66 mapped individuals, and the two sex-specific data sets comprised 29 individuals for each male and female. Although 37 females were available, we chose to use the same sample size as for males. The analysis with 37 females did not improve significantly eQTL detection, compared to the 29-female data set (data not shown). For eQTL detection, we used an interval-mapping approach (model 3 of the ZmapQTL module). Likelihood-ratio (LR) profiles [−2 ln(L0/L1)], representing the ratio of the likelihood of the null hypothesis (L0, no QTL in the marker interval) to the alternative hypothesis (L1, presence of a QTL in the marker interval) were generated for each transcript at every 2-cM interval on the linkage map of Rogers et al. (2007), using a window size of 10 cM. Three empirical thresholds for experimentwise type I error rates (0.1, 0.05, and 0.01) were determined for all significant eQTL (P < 0.05) by approximating their null distribution (H0). For each of them, we recorded the 10th-, 50th-, and 100th- ranked LR of 1000 random permutations. To achieve this, we developed a PYTHON language script for the UNIX version of QTL cartographer, which allowed the automatic running of iterations among traits. Users are first required to set up the analysis in a parameter file containing the list of traits to be analyzed, the number of iterations, the number of iterations for each trait (here 1000), and the percentile values to be sampled (e.g., 99th, 95th, and 90th). The PYTHON script records the parameter values and calls the different QTL cartographer programs, including Rmap (reads the map), Rcross (reads the cross), and Zmapqtl (runs association analysis between traits and markers). For each trait (transcript), both during and following each iteration, a function reads the output from Zmapqtl and records the GlobalMax values with each percentile value fixed. The script is available from the authors upon request.

Transcripts corresponded to EST library annotations of Von Schalburg et al. (2005). Those annotated as “unknown” were submitted to BLAST nucleotide and translated protein searches to determine if new gene identifications were possible. The highest BLAST score with a known function was recorded, for E ≤ 1 × 10−5, and successful transcript finds were annotated as “similar to.”

Functional annotation was done according to cGRASP. Functional categories include various biological processes (functions as annotated by cGRASP are in parentheses) and are defined as follows for the three most represented: (i) energy metabolism (electron transport, glycolysis, ATP synthesis, tricarboxylic acid cycle intermediate metabolism, tricarboxylic acid cycle, malate metabolism, hydrogen transport), (ii) muscle proteins (regulation of muscle contraction, muscle development, cytoskeleton organization), and (iii) protein synthesis (protein biosynthesis, electron transport, transcription, translational elongation, ribosome biogenesis). According to these criteria, we observed among the 2255 significantly expressed genes that 158 of them (7%) coded muscle proteins, 310 (14%) belonged to energetic metabolism, 484 (21%) participate in protein synthesis, and 946 (42%) were unknown functionally, according to the cGRASP annotation.

To illustrate the genomic distribution of eQTL, we divided the genome into 17.6-cM bins, corresponding to the average distance between markers for male and female maps (Rogers et al. 2007), which resulted in 366 bins. A Poisson distribution was assumed for calculating the probability of observing a given number of eQTL within any bin along the map. The mean of the Poisson distribution for each data set was estimated as the number of eQTL linkages/366 bins detected at α = 0.05. The proportion of positive vs. negative additive effects for eQTL (i.e., effect of segregating alleles on the amount of a given transcript) belonging to a given linkage group or eQTL hotspots (numerous eQTL located at the same bin) was tested using a Mann–Whitney test.

Expression profile differences between males and females from the backcross family progeny:

A classical paired analysis was conducted between 24 male and 24 female backcross progenies chosen among the 66 slides used for the eQTL mapping experiment This was the maximum number of slides available upon which a male was hybridized with a female. A dye swap between Cy3 and Alexa 647 fluorophores has been used because it minimizes bias inherent to the dyes' uneven fluorescence intensity (Churchill 2002). Transcription data were corrected for regional intensity-related bias, using an R-LOWESS algorithm, and analyzed by ANOVA, using the R/MAANOVA software package (Kerr et al. 2000) under a mixed-effect model [Yijgh = μ + Ai + Dj + Gg + Sh + (AD)ij + (GA)gi + (GD)gj + (GS)gh + epsilonijgh], where sex (S, male or female) and dye (D) were fixed, array (A) was random, and (epsilonijgh) was residual error. Since our analysis model was across genes, and because gene expression levels may have different variances (Cui et al. 2005; Kristiansson et al. 2006; Mathur and Dolo 2007), we used the shrinkage F statistic (FS) to detect significant differences (adjusted permuted P-values, α < 0.01) in level of gene transcription between males and females (Cui et al. 2005). As recommended by the authors, the FS null distribution was established by permutation analysis (1000 sample permutations), considering each array as a sampling unit.

Expression profile differences between pure dwarf and normal whitefish parental generations:

Dwarf and normal whitefish reared in a control environment were originally sampled in 1998 in Témiscouata Lake and Aylmer Lake (Quebec), respectively, and held in captivity at the Laboratoire Régional des Sciences Aquatiques (LARSA) facilities (Université Laval, Quebec). Family crosses were made in 2001 to generate pure F1 breeds. Both normal and dwarf lines were kept at all times under the same environmental conditions (water temperature, photoperiod, diet). In September 2004, 12 randomly selected individuals of comparable sizes for each ecotype were measured (mean fork length, dwarf, 22.9 cm, SD = 1.8 cm; normal, 28.9 cm, SD = 2.2 cm) and then killed with a 0.001% Eugenol solution just prior to tissue extraction. Muscle tissue samples were immediately frozen on dry ice and stored at −80°.

RNA extraction and microarray experiments were processed as described above. Two samples (one dwarf and one normal) were differentially colored by fluorescence (Cy3 and Alexa 647). The analysis focused on the same 2255 transcripts that were used for eQTL mapping. Before statistical analysis, data preparation was processed as described for the sex-paired analysis. The ANOVA was conducted under a mixed-effect model [Yijkg = μ + Ai + Dj + Tk + Gg + (AD)ij + (GA)gi + (GD)gj + (GT)gk + epsilonijkg], where sample (T, dwarf or normal), gene (G), and dye (D) were fixed, array (A) was random, and (epsilonijkg) was residual error. A permutation-based F-test (FS, 1000 sample permutations) was used to detect significant differences (permuted P-values, α < 0.05) in level of gene transcription between dwarf and normal whitefish. Changes in transcription levels were calculated as a D/N ratio, where the mean transcription in dwarf individuals (D) was divided by mean transcription in normal individuals (N) for each gene.

RESULTS

Detection of eQTL:

Among the 66 backcross progeny examined, a total of 2255 genes were significantly transcribed (see materials and methods). Striking differences in terms of numbers and effects of detected eQTL were observed when using the combined (a data set of 66 mapped individuals, analyzed with an ANOVA model discarding the sex effect) and the two sex-specific (29 individuals for each male and female) data sets. Thus, the proportion of variance explained (PVE) by individual eQTL was much higher for both males (median value of PVE = 66%; range, 47–81%) and females (PVE = 73%; range, 50–80%) than for the combined data set (PVE = 30%; range, 24–75%), which suggested a potential Beavis effect (Lynch and Walsh 1998) and/or the conflict between sexes in the combined data set (supplemental Table S1, Figure 1). Moreover, for any given level of stringency, the number of eQTL detected in the female data set was significantly higher than that in both the male and the combined data sets (Table 1). Similarly, the number of eQTL detected in the male data set was always higher than that detected in the combined data set for any threshold value. This indicated that reduced sample size in the sex-specific data sets did not create a downward bias in eQTL detection and that contrasting patterns between sexes blurred eQTL detection when using the combined data set. Furthermore, the three data sets were also contrasted in terms of absolute values of additive effects (i.e., amount of a given transcript modulated by segregation of alleles between dwarf and normal genomes). The female data set exhibited significantly more eQTL with absolute additive effects ≥1.00 (77/174) when compared to the male data set (3/55, P-value <0.001) and to the combined data set (1/30, P-value <0.0001). The mean absolute values of additive effects were also significantly higher in the female data set (mean = 0.99), when compared to the male (mean = 0.64, P-value = 0) and combined (mean = 0.58, P-value = 0) data sets.

TABLE 1
Number of eQTL detected for a given empirical threshold for experimentwise type I error rates of 0.10, 0.05, and 0.01 that were determined for each significantly mapped transcript (P < 0.05) by recording the 10th-, 50th-, and 100th-ranked LRs ...
Figure 1.
Distribution of eQTL according to their proportion of variation explained (PVE) for all 2255 expressed genes. Female-specific (a), male-specific (b), and combined (c) data sets are shown.

Differences between the combined and sex-specific data sets were also found in terms of number of eQTL associated with different gene transcripts. Thus, at the permuted genomewide α of 0.10 for suggestive eQTL detection, the average number of eQTL associated with a given transcript was comparable for both the combined (1.085) and male (1.190) data sets but substantially higher for the female data set (2.747) (Fisher's test, P < 0.0001) (Figure 2). Thus, no more than two and three significant eQTL per transcript were found in the combined and male data sets, respectively, while up to eight significant eQTL per transcript were found in the female data set. The same differential pattern was still significant at a permuted genomewide α of 0.05, but not at α = 0.01. This supports the observation that the trans-acting effects are generally of smaller effect than cis-acting effects and will not be detectable at overly conservative thresholds (Gibson and Weir 2005; Petretto et al. 2006). Finally, there were only three transcripts in common between male and female data sets, but with distinct eQTL locations: a nucleolar RNA helicase II (LG32f in male and LG19f in female), a 40S ribosomal protein S2 (LG25m in male and LG35m in female), and an ornithine decarboxylase antizyme (LG32f in male and LG14f in female). This suggests that transcript regulation was itself strongly influenced by sex.

Figure 2.
Number of eQTL detected per transcript. Results are shown for combined (solid), male (striped), and female (shaded) data sets. The x-axis: classes of eQTL number per transcript; y-axis: class sizes.

Functional groups:

At genomewide α of 0.05, 3 significant eQTL (8%) were detected, which belonged to energetic metabolism in the combined data set, 4 (11%) belonged to muscle proteins, and 13 (34%) belonged to protein synthesis functional groups. In the male-specific data set, 8 eQTL (15%) belonged to energetic metabolism, 7 (13%) belonged to muscle proteins, and 14 (26%) belonged to protein synthesis. In the female-specific data set, 23 eQTL (13%) belonged to energetic metabolism, 16 (9%) belonged to muscle proteins, and 30 (17%) belonged to protein synthesis and cell growth (Table 2, supplemental Table S1). These proportions of functional categories did not depart significantly from what was observed among the 2255 significantly expressed genes (see materials and methods). Significant directionality of additive effect for two functional groups, energetic metabolism and/or muscle contraction regulation, was observed in the three data sets (Table 2). Energetic metabolism exhibited significant directionality in both the combined and the male data sets, but with opposite effects. Then, muscle contraction proteins exhibited significant directionality in both male and female data sets, but here again, with opposite effects: positive in males, negative in females.

TABLE 2
Summary of directionality of additive changes sorted by functional groups

Nonrandom chromosomal distribution of eQTL:

The eQTL linkages were not evenly distributed across the genome, particularly so for sex-specific data sets (Figure 3). Statistical significance of nonrandom distribution was assessed for a genomewide α of 0.05, for which totals of 33, 54, and 174 eQTL were detected for the combined, male, and female data sets, respectively (Table 1). For each of the data sets, we conservatively considered the occurrence of at least 4 eQTL within a given bin as an indication of eQTL hotspots. Indeed, the probability of observing ≥4 occurrences of eQTL in any given bin was <0.001 for the female data set when assuming a Poisson distribution. According to this criterion, both male and female data sets (but not the combined data set) revealed hotspots. In females, two major hotspots were located on the LG 25m subtelomeric region, with 53 eQTL located at position 229 and 34 eQTL at position193, both of which exhibited a significant mean directional additive effect (Figure 2, Table 3). Furthermore, absolute values of additive effects exceeded 1.00 for 77% of these 87 eQTL, whereas this proportion fell to 12% for the 79 remaining eQTL mapped outside the LG25m. This suggests that these particular chromosomal regions, both located on the same linkage group, have the most pervasive influence on transcript regulation and that this influence is strongly sex biased. Moreover, when considering functional groups, transcripts involved in protein synthesis, cell growth, and division exhibited systematically opposed additive effects to those associated with energy metabolism, proteolysis, muscle contraction, and repair. In this study, the additive effect of an eQTL represents the influence of a normal allele segregating in a dwarf genetic background on the expression regulation of a particular gene. For example, a positive additive effect means that the normal allele upregulated the transcription of a given gene. According to this, six transcripts involved in protein synthesis and cell growth, including an histone H1 homolog to protein, a ribosomal biogenesis NEP1 protein, a highly conserved transcript in eukaryotes and archaea identified as an essential growth factor regulator (Eschrich et al. 2005), a GDP-l-fucose synthetase, tightly involved in growth and development (Smith et al. 2002), and a RhoGTPase, which is involved in development of the striated muscles and in myofibrillogenesis in vertebrates (O'Brien et al. 2000), exhibited opposite additive effects to four enzymes involved in energetic metabolism (a triose phosphate isomerase, a sarcoplasmic/endoplasmic reticulum ATPase, a glycerol-3-phosphate dehydrogenase, and a kinase precursor) and to six muscle protein transcripts [two actins, two collagen α-precursors, and a myosin heavy chain (supplemental Table S1)]. In addition, there were three small hotspots encompassing respectively five, four, and six transcripts, which were located on LG19, LG25m, and LG35m, respectively. These harbored very few eQTL exceeding an additive effect of 1.00. Concerning the male data set, only one hotspot was detected on LG3f, located at position 62.89, featuring 4 eQTL: a ribosomal protein, an ATP synthase, a putative NADH, and an ubiquinol cytochrome c reductase, all of them exhibiting absolute values of additive effect <1.00.

Figure 3.
Genome map showing eQTL (α < 0.05) for the three data sets: combined, male, and female. The x-axes: numbers correspond to successively numbered markers on each linkage group where each interval corresponds to a 17.6-cM bin. Successive ...
TABLE 3
Hotspots detected among the three data sets

Differentially expressed genes between progeny of different sexes:

A total of 478 of the 2255 genes (21%) were differentially expressed between 24 male and 24 female progeny of the backcross family (adjusted permuted P-values, α < 0.01) (supplemental Table S2). Genes belonging to several functional groups were clearly transcribed differentially between sexes. Namely oxidative phosphorylation genes were upregulated in males, whereas glycolytic enzymes, muscle proteins, and ribosomal proteins were upregulated in females. However, only 20 (4%) differentially expressed genes between males and females were associated with eQTL segregating in the female data set. This suggests that regulation settings for 96% of the transcripts differentially expressed between sexes were not sensitive to the genetic background of the backcross progenies, in which each individual represents a unique combination of dwarf and normal alleles.

Differentially expressed genes between dwarf and normal parental species:

A total of 253 of the 2255 genes (11%) were differentially expressed between 12 pure dwarf and 12 normal whitefish (permuted P-values, α < 0.05) (supplemental Table S3). Of these, 34 transcripts were associated with 53 eQTL (20% of the transcripts segregating among the three data sets), 22 in the female data set (Table 4A), 12 in the male data set (Table 4B), and one in the combined data set. The differentially expressed transcripts segregating in male and female data sets belonged mainly to three functional groups: muscle contraction, energy and repair (parvalbumin, myosin, actin, troponin, β-thymosin, sarcoplasmic/endoplasmic ATPase, and creatine kinase), energetic metabolism (pyruvate kinase, ATPase subunits G and C, NADH dehydrogenase, and cytochrome c oxidase polypeptide III), and protein synthesis (ribosomal proteins and zinc finger protein). Genes belonging to muscle contraction, energy, and repair were systematically upregulated in dwarf, except for troponin, whereas both energetic metabolism and protein synthesis functional groups were not (Table 4B). Also, 125 transcripts for which regulating regions segregated in male- and female-specific data sets were not differentially expressed between pure dwarf and normal whitefish, which suggests they may be associated with either interindividual differences or stabilizing selection.

TABLE 4
Transcripts differentially expressed between pure dwarf and normal whitefish associated with eQTL

DISCUSSION

Documenting the genetic architecture of gene expression provides a strong means for investigating the importance of the transcriptomic regulation during the course of an adaptive radiation in whitefish. Here, this approach revealed a complex genetic architecture of white muscle transcription, a tissue for which functional differences associated with swimming activity, energy metabolism, and growth have been well documented in the adaptive divergence of dwarf and normal whitefish species pairs (Derome and Bernatchez 2006; Derome et al. 2006). The first striking result of this study was the much higher number of chromosomal regions influencing transcript abundance in the female compared to the male data set of identical sample size. The second was the identification of two major nonrandom aggregations of highly significantly mapped transcript regulation loci (eQTL hotspots) in the female data set. Those were responsible for significant directional patterns of additive effects, therefore identifying potential candidate genes whose transcription levels may have coevolved under directional selection (Orr 1998) between dwarf and normal whitefish. Moreover, patterns of eQTL distribution were strongly contrasted between male and females data sets. Indeed, the number and distribution of eQTL per transcript provided strong indirect evidence for both putative cis and trans regulation, with trans regulation being apparently more prevalent in females than in males. The differential genetic architecture between sexes was also supported by a significantly higher number of transcripts showing segregation of their regulating regions in females relative to males. Furthermore, 20% of the transcripts associated with eQTL were differentially expressed between pure dwarf and normal progeny, and those mostly belonged to muscle energy and repair, energetic metabolism, and protein synthesis, therefore identifying potential candidate genes involved in the adaptive and/or historical genetic divergence of dwarf and normal whitefish.

Sex-specific genetic architecture of gene regulation: Sex-specific effect or adaptive sex-linked effects?

As stated above, the most striking result of this study was the much higher number of chromosomal regions influencing transcript abundance in the female compared to the male data set of identical sample size. Although several studies recently reported sex-specific transcript regulation (Von Schalburg et al. 2005; McIntyre et al. 2006), including for muscle tissue (Yang et al. 2006), this possibility is challenged here by the alternative hypothesis that the sex-specific architecture we observed reflects adaptive sex-linked effects. In addition, other hypotheses concerning the genetic architecture of whitefish and the experimental cross itself have to be explored. Thus, it is well known that recombination is strongly repressed during male meiosis in salmonids (Sakamoto et al. 2000), with a female-to-male linkage map distance ratio of 6.4:1 reported in S. trutta (Gharbi et al. 2006). Such a discrepancy between male and female would significantly affect segregation of regulatory regions, which, in turn, could lead to differential eQTL detection between male- and female-specific data sets. Rogers et al. (2007) did not detect any significant sex differences between the recombination frequencies in the whitefish linkage map. They hypothesized that elevated recombination in female markers of hybrid origin was likely offset by suppressed recombination due to the heterogeneous genetic background of the female hybrids compared to the pure parent in the pedigree. This, in turn, would have resulted in comparable levels of recombination observed between sexes, in contrast with previous studies in salmonids. Despite this, we recorded significantly fewer eQTL in males but also in the whole data set compared to females. Moreover, despite numerous unique combinations between dwarf and normal alleles in the backcross family, >20% of expressed genes were differentially regulated between progeny of different sexes. Yet, 96% of these transcripts did not segregate in any of the three data sets, which suggests independence vis-à-vis their genetic background. Thus, most of the transcripts segregating in sex-specific data sets were not sex-biased genes. These transcripts were nonetheless under the control of the sex-determining region, which exhibited in that case a dependence vis-à-vis their genetic background. The fact that different transcripts segregated in either male or female data sets strongly suggests that the sex-determining region exerts pleiotropic effects. This in turn raises the hypothesis that pleiotropic effects of sex-determining regions may have played an important role in the adaptive divergence of dwarf and normal whitefish and may have contributed to their postzygotic reproductive isolation, as documented experimentally in previous studies comparing fitness of pure and hybrid crosses between dwarf and normal whitefish (Lu and Bernatchez 1998; Rogers and Bernatchez 2006).

Contrasting results among the combined and sex-specific data sets:

Another salient result of this study was that there was no eQTL co-occurrence among the combined and the two sex-specific data sets. While the combined data set represents segregation effects on transcript regulation devoid of sex-linked effect, male and female data sets likely reflected contrasted eQTL occurrence due to, everything else been equal, differential patterns of recombination in the sex-determining region. Sex chromosomes in salmonids are at the early stages of differentiation (Phillip et al. 2001), and male heterogamety and female homogamety are accepted as a general rule in salmonids (reviewed in Philipps and Rab 2001). Moreover, X and Y chromosomes still share a similar repertoire of functional genes, conserved across a wide range of salmonid species (Woram et al. 2003, Artieri et al. 2006). Also, male-specific loci of the sex-determining autosomal region of two comparable young heterogametic sex chromosome systems (the medaka, Oryzias latipes, and the stickleback, Gasterosteus aculeatus), were observed to not recombine and to include the sex-determining gene (Matsuda et al. 2002; Peichel et al. 2004).

Aside from this, both size and weight at the same age differed significantly between males and females of the backcross family (Psize = 0.018; Pweight < 1.10−6, one tailed t-test), associated with a variance four time lower in males, whereas these parameters did not vary between sexes either in dwarf or in normal whitefish from natural populations (data not shown). This suggests, first, that recombination between dwarf and normal genomes was responsible for both size and weight differences observed between sexes and, second, that these differences were mostly under the influence of a putative sex-determining region that did not recombine in males, as both weight and size variances were substantially reduced compared to females. This in turn suggests that dwarf Y-like-linked genes imposed hemizygous trans-acting effects among backcross males, despite autosomal dwarf/normal allele segregation. This hypothesis is reasonable because a majority of trans-regulation patterns occurred in the homogametic sex (see results) where segregation was theoretically possible along the whole sex-determining region. It is also consistent with the work of Wang et al. (2006), who observed a majority of sex-linked trans-regulation patterns in mouse. Taken together, it is likely that significantly lower eQTL occurrence in a male-specific sample would have resulted from a dominant effect of the dwarf Y-like nonrecombining autosomal region that would in turn impose strong pleiotropic effects in a 75% dwarf background genome of the backcross family. By contrast, free recombination between dwarf and normal X-like autosomal regions during meiosis in F1 females would have caused higher eQTL occurrence in the backcross female-specific data set. In other words, eQTL detected in the males would correspond to sex-linked effects outside the nonrecombining region of the Y-like autosomal region, whereas those detected in females would correspond mostly to sex-linked effects encompassing the whole X-like chromosomal region.

The allopatric origin of the two evolutionary lineages used to build the backcross family may also partly underlie this striking difference of eQTL abundance between male and female data sets. The dwarf and normal whitefish populations belonged to different evolutionary lineages, Acadian and Atlantic–Mississippian, issued from distinct glacial origins (see materials and methods). Therefore, 18,000–500,000 years of allopatric genetic divergence between glacial races (Bernatchez and Dodson 1990) were enforced with ~12,000 years of adaptive divergence that occurred between sympatric species and have potentially enhanced regional allelic incompatibilities between dwarf and normal genomes, perhaps also including the sex-determining region.

The major difference between males and females was further expressed in the nonrandom aggregation of eQTL at two tightly linked regulatory hotspots observed in females. Indeed, linkage group LG25m comprised >55% of all eQTL detected in the female data set, the majority (98%) of which were located at the distal tip of the linkage group. Furthermore, these two eQTL hotspots, which comprised 90% of all LG25m eQTL, exhibited a strikingly higher proportion of strong additive and directional effects, compared to eQTL found elsewhere in the genome (see results). Indeed, the mean absolute value of additive effects reached 1.22 (range = 0.52–1.87), which is the very upper range of additive effects observed in mouse (Petretto et al. 2006). Moreover, additive effects were significantly directional when considering eQTL belonging to separate functional groups, namely energy metabolism (P = 0.016), muscle proteins (P = 0.004), and protein synthesis (P = 0.019). Directional predominance of additive effects in one direction for the loci underlying the quantitative variation of a given trait provides strong indirect evidence for the role of directional selection in shaping genome architecture (Orr 1998). Altogether, when considering a functional group as a trait, this suggests that directional selection has been acting on patterns of gene regulation associated with these two tightly linked chromosomal regions. Strikingly, directionality of additive effects was systematically opposed between transcripts involved in different functional groups. Namely, normal whitefish alleles induced negative regulation effects on energetic metabolism and muscle proteins. On the contrary, normal whitefish alleles induced positive regulation effects on protein synthesis and cell growth functional groups.

This result is very consistent with both previous bioenergetic and population genomic studies on whitefish ecotype differences. First, Trudel et al. (2001) showed that dwarf ecotypes have a higher metabolic rate, partly associated with the cost of swimming, and a lower bioenergetic conversion efficiency, both of them associated with slower growth and younger age at sexual maturity. In addition, parallelism in gene transcription of the same tissue in natural population pairs of whitefish ecotypes showed that adaptive transcriptional divergence involved mainly genes associated with swimming activity and energy metabolism (Derome and Bernatchez 2006; Derome et al. 2006). Such a concentration of eQTL belonging to diverse functional groups at these too subtelomeric positions (193 and 229) suggests that two single master regulators located at these positions are likely to control their expression. Moreover, all transcripts regulated at position 193 appeared also regulated on position 229. This raises the hypothesis that two duplicated regulating sequences are acting on the same suite of genes. This hypothesis is reasonable as positions 193 and 229 are separated by two markers that are associated with two regulating regions of, respectively, one and four different transcripts.

Among the transcripts that belong to both eQTL hotspots, there are four genes that are known to play fundamental roles in transcription regulation. First, the DPY-30-like protein gene encodes for a highly conserved protein throughout evolution, which is the upstream regulator of dosage compensation machinery of X chromosomes (Dong et al. 2005). In addition, this protein plays an essential role in sex determination (Hsu and Meyer 1993) and may direct various developmental processes via epigenetic control (Jenuwein and Allis 2001). Furthermore, it has been recently demonstrated in Caenorhabditis elegans that at least two separated DNA motifs, termed Rex1 and Rex2, act in combination or separately to regulate dosage compensation machinery via two DPY variants' activation. Also, Rex1 and Rex2 are, respectively, located close to one of the two distinct DPY sequences (Mcdonel et al. 2006). Such genetic architecture of DPY regulation could be consistent with the observation of both its tightly linked eQTL located on LG25m. Therefore, DPY-30-like protein appears to be a very relevant candidate as a master regulating locus for both LG25m hotspots. Another such candidate is the Histone H1 homolog protamine transcript, which is known to play a pervasive role in transcriptome regulation (Bustin et al. 2005). Interestingly, Nagy et al. (2001) have demonstrated an interaction between a DPY-30 homolog and a histone protein as a sex-specific epigenetic regulation in Drosophila, where histone methylation acts on chromatin dynamics. Therefore, as histone methylation is conserved from yeast to humans (Strahl et al. 1999), the coregulation of both DPY-30-like protein gene and histone H1 in whitefish could potentially account for regulating the suite of genes mapped on two tightly linked subtelomeric positions of a single chromosome. Another transcript potentially involved in the regulation of the two hotspots is the small ubiquitin-like modifier (SUMO). SUMO has emerged as an important regulator of diverse pathways and activities, including protein localization and transcriptional regulation (Hamard et al. 2007), and was observed to interact with RAD23 (Ortolan et al. 2000), an UV excision repair protein that is itself involved in chromatin-mediated repression of transcription, and is alleviated by both histone modifications and ATP-dependent chromatin-remodeling activities (Gong et al. 2006). Overall, these results exemplify how natural selection could act on a single master regulator controlling expression of a suite of numerous genes belonging to diverse functional groups in a tissue highly involved in the adaptive divergence of whitefish and that would be ultimately able to modulate notably the organism phenotype (Derome and Bernatchez 2006; Derome et al. 2006).

When extending the link between functional groups and directionality of additive effects to all linkage groups, only eQTL for transcripts belonging to muscle contraction proteins still have significant directionality, for which normal alleles induced a downregulation. However, the magnitude of their additive effects was weaker than that of those belonging to LG25m, suggesting a less profound effect in phenotypic differentiation between dwarf and normal whitefish. Significant directionality of additive effects on muscle contraction proteins and energy metabolism was observed in the male data set as well but with both opposite and smaller effects than those observed in females. Thus, sex differences of functional-specific directionality of transcript regulation add further support to the hypothesis that natural selection acted on a genetic architecture of transcript regulation mostly under the influence of the sex-determining region.

The eQTL hotspots on the LG25m are particularly striking because no such eQTL concentration was observed elsewhere on the map, and especially not on the LG25f, where alleles issued from the grandmother segregated. This suggests that this hotspot of transcriptomic segregation was specific to grandfather alleles, and, everything else been equal, these alleles would be linked to the X-like region homologous to the Y-like nonrecombining region. Indeed, loci associated with the sex-determining region in stickleback were reported to segregate only in male maps (Peichel et al. 2004). However, although many sex effects mapped to the LG25m, including a reproductive timing pQTL (Rogers and Bernatchez 2007), two eQTL of DPY-30-like protein, from which both regulating sequences and the gene itself were located on the X chromosome in C. elegans (Mcdonel et al. 2006), the genetic map was not dense enough to unequivocally determine a sex-determining locus.

Overall, these contrasting results between sexes revealed that two major tightly linked master regulating regions that influence expression of 34 common genes belonging to functional groups are likely involved in the adaptive divergence between dwarf and normal whitefish. Also, the additive effects of normal inherited alleles into the dwarf genetic background illustrate the trade-off in energy allocation in these species pairs in accordance with previous bioenergetic and transcriptomic studies (Trudel et al. 2001; Derome and Bernatchez 2006; Derome et al. 2006; St-Cyr et al. 2008). As such, this work provides further insights into the genetic architecture of transcriptional regulation in the time course of an adaptive radiation, in which historical isolation during an allopatric phase followed as well by ecological interactions encountered in sympatry following secondary contact have been acting. Our results also add further support to the view that trans-acting factors play an important role in the evolution of gene expression (Wang et al. 2006), especially during the early stages of adaptive divergence.

Comparison between eQTL and transcripts differentially expressed between pure dwarf and normal progenies:

The eQTL detection is dependent on alleles with significant regulatory effects segregating between two genomes, therefore encompassing both interindividual genetic differences and eventually regulatory changes differentiating distinct evolutionary lineages that may or may not be under selective effect. This study revealed that sex-linked pleiotropic effects have played pervasive effects in the transcript segregation. Here, there was no co-occurrence of eQTL between the three data sets, and only one transcript differentially expressed associated with one eQTL in the combined data set (i.e., exhibiting transcriptomic segregation devoid of sex-linked effect). This supports further the importance of the sex-determining region in the genetic architecture differentiating dwarf and normal whitefish. Furthermore, 12 candidate genes were differentially expressed in the same direction between the parental dwarf and normal whitefish populations kept in controlled conditions at LARSA and that were used for genetic mapping, as well as one or two additional natural sympatric pairs studied by Derome et al. (2006): parvalbumin, actin, sarcoplasmic/endoplasmic reticulum calcium, creatine kinase, ATP synthase g, pyruvate kinase, lysozyme, and ribosomal protein were upregulated in dwarf whitefish whereas cytochrome oxidase subunit 3, putative NADH dehydrogenase, troponin, and serine/threonine protein phosphatase were upregulated in normal whitefish. This co-occurrence of directional changes for these 12 specific transcripts, along with segregation of their respective regulating regions, further suggests that the regulation of these genes evolved under natural selection during the adaptive radiation of the whitefish species complex.

Concluding remarks:

This study exemplifies how the integration of genetic and transcriptomic data offers a strong means for dissecting the functional genomic response to selection by separating mapping family-specific effects from genetic factors under selection, potentially involved in the phenotypic divergence of natural populations. In particular, the genetic dissection of gene regulation architecture underlying phenotypic divergence between dwarf and normal whitefish provided new insights toward understanding the molecular mechanisms underlying their divergence and, ultimately, their reproductive isolation. First, the integration of a paired analysis comparing expression across sex and eQTL mapping of sex-specific data sets revealed that the sex-determining region would have imposed strong pleiotropic effects on adaptive traits, which could in turn at least partly account for their postzygotic reproductive isolation. Second, a higher number of eQTL were recorded in female progeny, for which two major hotspots of transcript regulation on two discrete positions on the LG25m encompassed more than half of all detected eQTL. This provides evidence for the existence of two master control loci modulating the same suite of genes, belonging to diverse functional groups involved in whitefish phenotypic divergence, including muscle contraction, energetic metabolism, and protein synthesis. The pronounced directionality and magnitude of additive effects of eQTL found in these chromosomal regions provided strong albeit indirect support for the role of historical selection in shaping patterns of transcription for this large suite of genes. Third, the transcripts associated with eQTL that were differentially expressed between pure dwarf and normal whitefish progeny added further insights into identifying candidate genes involved in the genetic divergence of these species. Overall, this study brings support to the hypothesis that natural selection may have acted on a few master regulating loci, which appeared to be under the influence of the sex-determining region. These master regulating loci, in turn, may have shaped the transcription profile at whole suites of genes through extensive pleiotropic effects and ultimately modulated phenotypic traits that underlay the adaptive divergence and speciation in dwarf and normal whitefish species pairs.

Acknowledgments

We thank Serge Higgins and all the LARSA team for their support during the rearing of the whitefish, J. St-Cyr for tissue sampling, and Jennifer Lake for the parental line microarrays. We also thank N. Isabelle, S. McCairns, and S. Renaut, for their constructive comments on earlier versions of the manuscript. We also acknowledge the kind collaboration of B. F. Koop and W. S. Davidson, both leaders of the Consortium for Genomic Research on All Salmoids Project (cGRASP). Finally, we thank Lauren McIntyre and two anonymous referees for their very constructive comments and helpful suggestions. This research was financially supported by a Natural Sciences and Engineering research Council of Canada grant to L.B.

Notes

Expression data from this article have been deposited with the NCBI/Gene Expression Omnibus under accession nos. GSE11378, GPL6260, GSM276833GSM276834, GSM276836GSM276842, GSM276995GSM277089, and GSM277155GSM277195.

References

  • Artieri, C. G., L. A. Mitchell, S. H. Ng, S. E. Parisotto, R. G. Danzmann et al., 2006. Identification of the sex-determining locus of Atlantic salmon (Salmo salar) on chromosome 2. Cytogenet. Genome Res. 112 152–159. [PubMed]
  • Basten, C. J., B. S. Weir and Z. Z.-B., 2002. QTL Cartographer. Program in Statistical Genetics, Bioinformatics Research Center, Department of Statistics, North Carolina State University, Raleigh, NC.
  • Bernatchez, L., 2004. Ecological theory of adaptive radiation: an empirical assessment from coregonine fishes (Salmoniformes), pp. 175–207 in Evolution Illuminated: Salmon and Their Relatives, edited by A. P. Hendry and S. C. Stearns. Oxford University Press, Oxford.
  • Bernatchez, L., and J. J. Dodson, 1990. Allopatric origin of sympatric populations of Lake whitefish (Coregonus clupeaformis) as revealed by mitochondrial-DNA restriction analysis. Evolution 44 1263–1271.
  • Bochdanovits, Z., H. Van Der Klis and G. De Jong, 2003. Covariation of larval gene expression and adult body size in natural populations of Drosophila melanogaster. Mol. Biol. Evol. 20 1760–1766. [PubMed]
  • Brem, R. B., G. Yvert, R. Clinton and L. Kruglyak, 2002. Genetic dissection of transcriptional regulation in budding yeast. Science 296 752–755. [PubMed]
  • Bustin, M., F. Catez and J. H. Lim, 2005. The dynamics of histone H1 function in chromatin. Mol. Cell 17 617–620. [PubMed]
  • Churchill, G. A., 2002. Fundamentals of experimental design for cDNA microarrays. Nat. Genet. 32 490–495. [PubMed]
  • Cui, X., and G. A. Churchill, 2003. Statistical tests for differential expression in cDNA microarray experiments. Genome Biol. 4 210. [PMC free article] [PubMed]
  • Cui, X., J. T. Hwang, J. Qiu, N. J. Blades and G. A. Churchill, 2005. Improved statistical tests for differential gene expression by shrinking variance components estimates. Biostatistics. 6 59–75. [PubMed]
  • Derome, N., and L. Bernatchez, 2006. The transcriptomics of ecological convergence between 2 limnetic coregonine fishes (Salmonidae). Mol. Biol. Evol. 23 2370–2378. [PubMed]
  • Derome, N., P. Duchesne and L. Bernatchez, 2006. Parallelism in gene transcription among sympatric lake whitefish (Coregonus clupeaformis Mitchill) ecotypes. Mol. Ecol. 15 1239–1250. [PubMed]
  • Dong, X., Y. Peng, Y. Peng, F. Xu, X. He et al., 2005. Characterization and crystallization of human DPY-30-like protein, an essential component of dosage compensation complex. Biochim. Biophys. Acta 1753 257–262. [PubMed]
  • Ellegren, H., and A. K. Fridolfsson, 2003. Sex-specific mutation rates in salmonid fish. J. Mol. Evol. 56 458–463. [PubMed]
  • Ellegren, H., and J. Parsh, 2007. The evolution of sex-biased genes and sex-biased gene expression. Nat. Rev. Genet. 8 689–698. [PubMed]
  • Eschrich, D., M. Buchhaupt, P. Kötter and K. D. Entian, 2005. Nep1p (Emg1p), a novel protein conserved in eukaryotes and archaea, is involved in ribosome biogenesis. Curr. Genet. 40 326–338. [PubMed]
  • Fitzpatrick, M. J., 2004. Pleiotropy and the genomic location of sexually selected genes. Am. Nat. 163 800–808. [PubMed]
  • Gharbi, K., A. Gautier, R. G. Danzmann, S. Gharbi, T. Sakamoto et al., 2006. A linkage map for brown trout (Salmo trutta): chromosome homeologies and comparative genome organization with other salmonid fish. Genetics 172 2405–2419. [PMC free article] [PubMed]
  • Gibson, G., 2002. Microarrays in ecology and evolution: a preview. Mol. Ecol. 11 17–24. [PubMed]
  • Gibson, G., and B. Weir, 2005. The quantitative genetics of transcription. Trends Genet. 21 616–623. [PubMed]
  • Gong, F., D. Fahy and M. J. Smerdon, 2006. Rad4-Rad23 interaction with SWI/SNF links ATP-dependent chromatin remodeling with nucleotide excision repair. Nat. Struct. Mol. Biol. 13 902–907. [PubMed]
  • Hamard, P. J., M. Boyer-Guittaut, B. Camuzeaux, D. Dujardin, C. Hauss et al., 2007. Sumoylation delays the ATF7 transcription factor subcellular localization and inhibits its transcriptional activity. Nucleic Acids Res. 35 1134–1144. [PMC free article] [PubMed]
  • Hoekstra, H. E., and J. A. Coyne, 2007. The locus of evolution: evo devo and the genetics of adaptation. Evolution 61 995–1016. [PubMed]
  • Hsu, D. R., and B. J. Meyer, 1993. X chromosome dosage compensation and its relationship to sex determination in C. elegans. Semin. Dev. Biol. 4 96–106.
  • Itoh, Y., E. Melamed, X. Yang, K. Kampf, S. Wang et al., 2007. Dosage compensation is less effective in birds than in mammals. J. Biol. 6 2. [PMC free article] [PubMed]
  • Jacob, F., and J. Monod, 1961. Genetic regulatory mechanisms in the synthesis of proteins. J. Mol. Biol. 13 318–356. [PubMed]
  • Jenuwein, T., and C. D. Allis, 2001. Translating the histone code. Science 293 1074–1080. [PubMed]
  • Jin, W., R. M. Riley, R. D. Wolfinger, K. P. White, G. Passador-Gurgel et al., 2002. The contributions of sex, genotype and age to transcriptional variance in Drosophila melanogaster. Nat. Genet. 29 389–395. [PubMed]
  • Kerr, M. K., M. Martin and G. A. Churchill, 2000. Analysis of variance for gene expression microarray data. J. Comput. Biol. 7 819–837. [PubMed]
  • King, C. C., and A. C. Wilson, 1975. Evolution at two levels in humans and chimpanzees. Science 188 107–116. [PubMed]
  • Kirst, M., C. J. Basten, A. A. Myburg, Z. B. Zeng and R. R. Sederoff, 2005. Genetic architecture of transcript-level variation in differentiating xylem of a eucalyptus hybrid. Genetics 169 2295–2303. [PMC free article] [PubMed]
  • Kristiansson, E., A. Sjögren, M. Rudemo and O. Nerman, 2006. Quality optimised analysis of general paired microarray experiments. Stat. Appl. Genet. Mol. Biol. 5Article 10. [PubMed]
  • Lander, E. S., P. Green, J. Abrahamson, A. Barlow, M. Daley et al., 1987. MAPMAKER, an interactive computer package for constructing primary genetic linkage maps of experimental and natural populations. Genomics 1 174–181. [PubMed]
  • Landry, L., W. F. Vincent and L. Bernatchez, 2007. Parallel evolution of lake whitefish dwarf ecotypes in association with limnological features of their adaptive landscape. J. Evol. Biol. 20 971–984. [PubMed]
  • Lu, G., and L. Bernatchez, 1998. Experimental evidence for reduced hybrid viability between dwarf and normal ecotypes of lake whitefish (Coregonus clupeaformis Mitchill). Proc. R. Soc. Lond. Ser. B 265 1025–1030.
  • Lu, G. Q., and L. Bernatchez, 1999. Correlated trophic specialization and genetic divergence in sympatric lake whitefish ecotypes (Coregonus clupeaformis): support for the ecological speciation hypothesis. Evolution 53 1491–1505.
  • Lu, G., D. J. Basley and L. Bernatchez, 2001. Contrasting patterns of mitochondrial DNA and microsatellite introgressive hybridization between lineages of lake whitefish (Coregonus clupeaformis); relevance for speciation. Mol. Ecol. 10 965–985. [PubMed]
  • Lynch, M., and B. Walsh, 1998. Genetics and Analysis of Quantitative Traits. Sinauer Associates, Sunderland, MA.
  • Mank, J. E., L. Hultin-Rosenberg, E. Axelsson and H. Ellegren, 2007. Rapid evolution of female-biased, but not male-biased, genes expressed in the avian brain. Mol. Biol. Evol. 24 2698–2706. [PubMed]
  • Marinotti, O., E. Calvo, Q. K. Nguyen, S. Sanayake, J. M. Ribeiro et al., 2006. Genome-wide analysis of gene expression in adult Anopheles gambiae. Insect Mol. Biol. 15 1–12. [PubMed]
  • Mathur, S., and S. Dolo, 2007. A new efficient statistical test for detecting variability in the gene expression data. Stat. Methods Med. Res. (in press). [PubMed]
  • Matsuda, M., Y. Nagahama, A. Shinomiya, T. Sato, C. Matsuda et al., 2002. DMY is a Y-specific DM-domain gene required for male development in the medaka fish. Nature 417 559–563. [PubMed]
  • Mcdonel, P., J. Jans, B. K. Peterson and J. Meyer, 2006. Clustered DNA motifs mark X chromosomes for repression by a dosage compensation complex. Nature 444 614–618. [PMC free article] [PubMed]
  • McIntyre, L. M., L. M. Bono, A. Genissel, R. Westerman, D. Junk et al., 2006. Sex-specific expression of alternative transcripts in Drosophila. Genome Biol. 7 R79. [PMC free article] [PubMed]
  • Mckay, T. F. C., 2001. The genetic architecture of quantitative traits. Annu. Rev. Genet. 35 303–339. [PubMed]
  • Morley, M., C. M. Molony, T. M. Weber, J. L. Devlin, K. G. Ewens et al., 2004. Genetic analysis of genome-wide variation in human gene expression. Nature 430 743–747. [PMC free article] [PubMed]
  • Nagy, P. L., J. Griesenbeck, R. D. Kornberg and M. L. Cleary, 2001. A trithorax-group complex purified from Saccharomyces cerevisiae is required for methylation of histone H3. Proc. Natl. Acad. Sci. USA 99 90–94. [PMC free article] [PubMed]
  • O'Brien, S. P., K. Seipel, Q. G. Medley, R. Bronson, R. Segal et al., 2000. Skeletal muscle deformity and neuronal disorder in Trio exchange factor-deficient mouse embryos. Proc. Natl. Acad. Sci. USA 97 12074–12078. [PMC free article] [PubMed]
  • Oleksiak, M. F., G. A. Churchill and D. L. Crawford, 2002. Variation in gene expression within and among natural populations. Nat. Genet. 32 261–266. [PubMed]
  • Oleksiak, M. F., J. L. Roach and D. L. Crawford, 2005. Natural variation in cardiac metabolism and gene expression in Fundulus heteroclitus. Nat. Genet. 37 67–72. [PMC free article] [PubMed]
  • Orr, H. A., 1998. Testing natural selection vs. genetic drift in phenotypic evolution using quantitative trait locus data. Genetics 149 2099–2104. [PMC free article] [PubMed]
  • Ortolan, T. G., P. Tongaonkar, D. Lambertson, L. Chen, C. Schauber et al., 2000. The DNA repair protein rad23 is a negative regulator of multi-ubiquitin chain assembly. Nat. Cell. Biol. 2 601–608. [PubMed]
  • Parisi, M., R. Nuttall, P. Edwards, J. Minor, D. Naiman et al., 2004. A survey of ovary-, testis-, and soma-biased gene expression in Drosophila melanogaster adults. Genome Biol. 5 R40. [PMC free article] [PubMed]
  • Parker, G. A., and L. Partridge, 1998. Sexual conflict and speciation. Philos. Trans. R. Soc. Lond. B Biol. Sci. 353 261–274. [PMC free article] [PubMed]
  • Peichel, C. L., J. A. Ross, C. K. Matson, M. Dickson, J. Grimwood et al., 2004. The master sex-determination locus in threespine sticklebacks is on a nascent Y chromosome. Curr. Biol. 14 1416–1424. [PubMed]
  • Petretto, E., J. Mangion, N. J. Dickens, S. A. Cook, M. K. Kumaran et al., 2006. Heritability and tissue specificity of expression quantitative trait loci. PLoS Genet. 20 e172. [PMC free article] [PubMed]
  • Philipps, R. B., and P. Rab, 2001. Chromosome evolution in the Salmonidae (Pisces): an update. Biol. Rev. 76 1–25. [PubMed]
  • Phillip, R. B., N. R. Konkol, K. M. Reed and J. D. Stein, 2001. Chromosome painting supports lack of homology among sex chromosomes in Oncorhynchus, Salmo, and Salvelinus (Salmonidae). Genetica 11 119–123. [PubMed]
  • Pröschel, M., Z. Zang and J. Parsh, 2006. Widespread adaptive evolution of Drosophila genes with sex-biased expression. Genetics 174 893–900. [PMC free article] [PubMed]
  • Prud'homme, B., N. Gompel and S. B. Carroll, 2007. Emerging principles of regulatory evolution. Proc. Natl. Acad. Sci. USA 104 8605–8612. [PMC free article] [PubMed]
  • Purugganan, M., 1998. The molecular evolution of development. BioEssays 20 700–711. [PubMed]
  • Rantz, J. M., C. I. Castillo-Davis, C. D. Meiklejohn and D. L. Hartl, 2003. Sex-dependent gene expression and evolution of the Drosophila transcriptome. Science 300 1742. [PubMed]
  • Reeves, H. K., and P. W. Sherman, 1993. Adaptation and the goals of evolutionary research. Q. Rev. Biol. 68 1.
  • Richards, S., Y. Liu, B. R. Bettencourt, P. Hradecky, S. Letovsky et al., 2005. Comparative genome sequencing of Drosophila pseudoobscura: chromosomal, gene, and cis-element evolution. Genome Res. 15 1–18. [PMC free article] [PubMed]
  • Rise, M. L., K. R. Von Schalburg, G. A. Cooper and B. F. Koop, 2007. Salmonid DNA microarrays and other tools for functional genomics research, pp. 369–411 in Aquaculture Genome Technologies, edited by Z. Liu. Blackwell Publishing, Oxford.
  • Roberge, C., H. Guderley and L. Bernatchez, 2007. Genome-wide identification of genes under selection: gene transcription Qst scan in diverging Atlantic salmon subpopulations. Genetics 177 1011–1022. [PMC free article] [PubMed]
  • Rogers, S. M., and L. Bernatchez, 2005. Integrating QTL mapping and genome scans towards the characterization of candidate loci under parallel selection in the lake whitefish (Coregonus clupeaformis). Mol. Ecol. 14 351–361. [PubMed]
  • Rogers, S. M., and L. Bernatchez, 2006. The genetic basis of intrinsic and extrinsic postzygotic reproductive isolation jointly promoting speciation in the lake whitefish species complex (Coregonus clupeaformis). J. Evol. Biol. 19 1979–1994. [PubMed]
  • Rogers, S. M., and L. Bernatchez, 2007. The genetic architecture of ecological speciation and the association with signatures of selection in natural lake whitefish (Coregonus sp. Salmonidae) species pairs. Mol. Biol. Evol. 24 1423–1438. [PubMed]
  • Rogers, S. M., V. Gagnon and L. Bernatchez, 2002. Genetically based phenotype-environment association for swimming behaviour in lake whitefish ecotypes (Coregonus clupeaformis Mitchill). Evolution 56 2322–2329. [PubMed]
  • Rogers, S. M., N. Isabelle and L. Bernatchez, 2007. Linkage maps of the dwarf and normal lake whitefish (Coregonus clupeaformis) species and their hybrids reveal the genetic architecture of population divergence. Genetics 175 375–398. [PMC free article] [PubMed]
  • Sakamoto, T., R. G. Danzmann, K. Gharbi, P. Howard, A. Ozaki et al., 2000. A microsatellite linkage map of rainbow trout (Oncorhynchus mykiss) characterized by large sex-specific differences in recombination rates. Genetics 155 1331–1345. [PMC free article] [PubMed]
  • Schadt, E. E., S. A. Monks, T. A. Drake, A. J. Lusis and N. Che et al., 2003. Genetics of gene expression surveyed in maize, mouse and man. Nature 422 297–302. [PubMed]
  • Servedio, M. R., and G. P. Saetre, 2006. Speciation as a positive feedback loop between postzygotic and prezygotic barriers to gene flow. Proc. Biol. Sci. 270 1473–1479. [PMC free article] [PubMed]
  • Singh, R. S., and R. J. Kulathinal, 2000. Sex gene pool evolution and speciation: a new paradigm. Genes Genet. Syst. 75 119–130. [PubMed]
  • Smith, P. L., J. T. Myers, C. E. Rogers, L. Zhou, B. Petryniak et al., 2002. Conditional control of selectin ligand expression and global fucosylation events in mice with a targeted mutation at the FX locus. J. Cell Biol. 158 801–815. [PMC free article] [PubMed]
  • St-Cyr, J., N. Derome and L. Bernatchez, 2008. The transcriptomics of life-history trade-offs in whitefish species pairs (Coregonus sp.). Mol. Ecol. 17: (in press). [PubMed]
  • Strahl, B. D., R. Ohba, R. G. Cook and C. D. Allis, 1999. Methylation of histone H3 at lysine 4 is highly conserved and correlates with transcriptionally active nuclei in Tetrahymena. Proc. Natl. Acad. Sci. USA 96 14967–14972. [PMC free article] [PubMed]
  • Trudel, M., A. Tremblay, R. Schetagne and J. B. Rasmussen, 2001. Why are dwarf fish so small? An energetic analysis of polymorphism in lake whitefish (Coregonus clupeaformis). Can. J. Fish. Aquat. Sci. 58 394–405.
  • Tusher, V. G., R. Tibshirani and G. Chu, 2001. Significance analysis of microarrays applied to the ionizing radiation response. Proc. Natl. Acad. Sci. USA 98 5116–5121. [PMC free article] [PubMed]
  • Von Schalburg, K. R., M. L. Rise, G. D. Brown, W. S. Davidson and B. F. Koop, 2005. A comprehensive survey of the genes involved in maturation and development of the rainbow trout ovary. Biol. Reprod. 72 687–699. [PubMed]
  • Vuylsteke, M., H. Daele, A. Vercauteren, M. Zabeau and M. Kuiper, 2006. Genetic dissection of transcriptional regulation by cDNA-AFLP. Plant J. 45 439–446. [PubMed]
  • Wang, P. J., J. R. McCarrey, F. Yang and D. C. Page, 2001. An abundance of X-linked genes expressed in spermatogonia. Nat. Genet. 4 422–426. [PubMed]
  • Wang, S., N. Yehya, E. E. Schadt, H. Wang, T. A. Drake et al., 2006. Genetic and genomic analysis of a fat mass trait with complex inheritance reveals marked sex specificity. PLoS Genet. 2 e5. [PMC free article] [PubMed]
  • Wayne, M., and L. McIntyre, 2002. Combining mapping and arraying: an approach to candidate gene identification. Proc. Natl. Acad. Sci. USA 99 14903–14906. [PMC free article] [PubMed]
  • Wayne, M. L., M. Telonis-Scott, L. M. Bono, L. Harshman, A. Kopp et al., 2007. Simpler mode of inheritance of transcriptional variation in male Drosophila melanogaster. Proc. Natl. Acad. Sci. USA 104 18577–18582. [PMC free article] [PubMed]
  • Whitehead, A., and D. L. Crawford, 2005. Variation in tissue-specific gene expression among natural populations. Genome Biol. 6 R13. [PMC free article] [PubMed]
  • Woram, R. A., K. Gharbi, T. Sakamoto, B. Hoyheim, L. E. Holm et al., 2003. Comparative genome analysis of the primary sex-determining locus in salmonid fishes. Genome Res. 13 272–280. [PMC free article] [PubMed]
  • Yang, X., E. E. Schadt, S. Wang, H. Wang, A. P. Arnold et al., 2006. Tissue-specific expression and regulation of sexually dimorphic genes in mice. Genome Res. 16 995–1004. [PMC free article] [PubMed]
  • Yang, Y. J., and T. Speed, 2002. Design issues for cDNA microarray experiments. Nat. Rev. Genet. 3 579–588. [PubMed]
  • Zang, Z., T. M. Hambuch and J. Parsch, 2004. Molecular evolution of sex-biased genes in Drosophila. Mol. Biol. Evol. 21 2130–2139. [PubMed]

Articles from Genetics are provided here courtesy of Genetics Society of America

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...