• Loading metrics

High-Resolution Mapping of Crossover and Non-crossover Recombination Events by Whole-Genome Re-sequencing of an Avian Pedigree

  • Linnéa Smeds ,

    Contributed equally to this work with: Linnéa Smeds, Carina F. Mugal

    Affiliation Department of Evolutionary Biology, Evolutionary Biology Centre, Uppsala University, Uppsala, Sweden

  • Carina F. Mugal ,

    Contributed equally to this work with: Linnéa Smeds, Carina F. Mugal

    Affiliation Department of Evolutionary Biology, Evolutionary Biology Centre, Uppsala University, Uppsala, Sweden

  • Anna Qvarnström,

    Affiliation Department of Animal Ecology, Evolutionary Biology Centre, Uppsala University, Uppsala, Sweden

  • Hans Ellegren

    Affiliation Department of Evolutionary Biology, Evolutionary Biology Centre, Uppsala University, Uppsala, Sweden

High-Resolution Mapping of Crossover and Non-crossover Recombination Events by Whole-Genome Re-sequencing of an Avian Pedigree

  • Linnéa Smeds, 
  • Carina F. Mugal, 
  • Anna Qvarnström, 
  • Hans Ellegren


Recombination is an engine of genetic diversity and therefore constitutes a key process in evolutionary biology and genetics. While the outcome of crossover recombination can readily be detected as shuffled alleles by following the inheritance of markers in pedigreed families, the more precise location of both crossover and non-crossover recombination events has been difficult to pinpoint. As a consequence, we lack a detailed portrait of the recombination landscape for most organisms and knowledge on how this landscape impacts on sequence evolution at a local scale. To localize recombination events with high resolution in an avian system, we performed whole-genome re-sequencing at high coverage of a complete three-generation collared flycatcher pedigree. We identified 325 crossovers at a median resolution of 1.4 kb, with 86% of the events localized to <10 kb intervals. Observed crossover rates were in excellent agreement with data from linkage mapping, were 52% higher in male (3.56 cM/Mb) than in female meiosis (2.28 cM/Mb), and increased towards chromosome ends in male but not female meiosis. Crossover events were non-randomly distributed in the genome with several distinct hot-spots and a concentration to genic regions, with the highest density in promoters and CpG islands. We further identified 267 non-crossovers, whose location was significantly associated with crossover locations. We detected a significant transmission bias (0.18) in favour of ‘strong’ (G, C) over ‘weak’ (A, T) alleles at non-crossover events, providing direct evidence for the process of GC-biased gene conversion in an avian system. The approach taken in this study should be applicable to any species and would thereby help to provide a more comprehensive portray of the recombination landscape across organism groups.

Author Summary

Homologous chromosomes exchange genetic material during cell division at meiosis by the process of crossover recombination. Although such crossover events are visually identifiable by cytogenetic techniques, it has remained a challenge to pinpoint the location of crossovers at the DNA sequence level. An emerging novel possibility to approach this challenge is to exploit the high resolution offered by re-sequencing of multiple individuals in species in which a genome assembly is available. Specifically, by sequencing members of a family, the inheritance of chromosomal segments can be followed and the location of crossover as well as non-crossover recombination determined. We performed such an endeavour in the collared flycatcher, a songbird species that has been in focus for extensive ecological and evolutionary research. We found that crossover events were concentrated to certain ‘hot-spot’ regions, and that the density of such events was highest in and close to genes. A higher rate of crossover recombination was found in males than in females and in males, but not in females, the rate of crossover increased towards ends of chromosomes. The location of non-crossovers was significantly associated with that of crossovers. We could further document an unbalanced transmission of genetic variants at non-crossover events via to the process of GC-biased gene conversion.


Meiotic recombination is intimately related to the evolution of sexual reproduction. It occurs early in meiosis and is commonly initiated by double-strand breaks (DSBs) that are catalysed by the SPO11 protein [1]. The broken ends are processed and their repair can either lead to crossovers (COs), which involve an exchange of chromatid arms and assist the proper segregation of homologous chromosomes during meiosis I, or non-crossovers (NCOs), i.e. recombination events without an exchange of chromatid arms. We here use the term ‘recombination’ to collectively refer to both types of events, while CO and NCO are used to refer to the respective outcome of recombination.

COs are critical to several evolutionary processes [2], such as the efficacy of selection (Hill-Robertson interference; [3]), and the evolution of sex chromosomes [4]. COs further modulate variation in levels of nucleotide diversity along chromosomes [58] and genetic differentiation between populations and species [9], and, together with selection, govern the character and extent of linkage disequilibrium [10]. Moreover, in addition to breaking up linkage and re-shuffling alleles, recombination affects the evolution of base composition via GC-biased gene conversion (gBGC) [2, 1113]. gBGC is a process that leads to a preferential transmission of GC-alleles over AT-alleles close to recombination-initiating DSBs. Base pair mismatches establish in heteroduplex DNA, which is formed as part of the repair pathway of DSBs, whenever homologous chromosomes carry different alleles. The transmission bias arises because mismatches that result from AT/GC heterozygous sites are resolved in favour of G:C base pairs.

CO rates (often measured as cM/Mb) can be estimated by combining data on CO fractions between markers in linkage analyses and physical information on the location of markers in the genome. Typically, resolution is limited by the density of available markers for genotyping and the number of meiosis in which the segregation of markers from parents to offspring can be followed. Recent development of arrays with tens or even hundreds of thousands of single nucleotide polymorphism (SNP) markers have offered increased resolution [1416] but the number of genotypes required to identify CO events between closely located markers still represents a limiting factor for fine-scale assessment of CO rates in most non-model organisms. Yet, comparisons of linkage maps and genome sequences [17, 18] have improved our understanding of the broad-scale patterns of CO rate variation concerning, for example, rate differences between species [1921], chromosomes [22], and sexes [23], as well as regional heterogeneity along chromosomes [24, 25].

The application of whole-genome re-sequencing to population genomics provides an indirect means to the estimation of fine-scale CO rates and can allow localization of historical CO events [26]. Specifically, estimated levels of linkage disequilibrium (LD) between pairs of segregating sites along chromosomes can be transformed into the scaled population recombination parameter (ρ = 4Ner), which can be used as a proxy for CO rate; high levels of LD are indicative of low CO rates, while low levels of LD are most easily explained by a high rate of COs. However, a drawback of this approach is that LD can be influenced by other forces than CO, such as selection, population structure and migration. Moreover, patterns of LD are the result of historical processes and do not necessarily reflect the properties of contemporary CO [27].

Whole-genome re-sequencing can also be used to get direct estimates of recombination rates. Specifically, re-sequencing of crosses [2831], pedigrees [32, 33], sperm and oocytes [3437] or spores [28] provide new and exciting direct approaches for localization of recombination events at high resolution and for the estimation of recombination rates. In principle, the density of informative polymorphisms that distinguish homologous chromosomes determines the resolution. Sequencing of gametes provides the structure of new haplotypes formed after CO events and meiotic tetrad analysis is particularly attractive in this respect since all four products from a single meiosis can be recovered and characterized [38]. It allows not only identifying CO events at high resolution but 3:1 inheritance between sister gametes implies that NCO gene conversion events can also be traced [28]. In organisms in which tetrad analysis is not possible, phased sequencing data from pedigreed individuals, most easily obtained if three generations can be followed, is technically less demanding to generate than comparable data from single-cell analysis.

One important conclusion from studies on recombination rate variation at high resolution is the realization that recombination events are often concentrated to specific genomic regions, so-called hot-spots. These are likely to coincide with regions accessible for DSB formation. In budding yeast (Saccharomyces cerevisiae), nucleosome occupancy and the histone H3 lysine 4 trimethylation (H3K4me3) chromatin modification facilitate the formation of DSBs by changing the accessibility for SPO11, and high rates of DSBs are observed in close proximity to transcription start sites (TSSs) [39, 40]. A similar picture is seen in plants; high rates of recombination are observed in close proximity of TSSs but also in close proximity of transcription termination sites (TTSs) [4143]. In humans and some other mammals (but not all [44]), the localization of hot-spots is associated with certain sequence motifs that are recognized by PRDM9 [4548], a zinc finger protein that trimethylates H3K4me3 [49]. In this case, PRDM9 binding occurs mainly in intergenic regions, and within genes, with a lowered rate of recombination close to TSSs [5052]. While the co-localization of recombination and transcription initiation seems to be a widespread and likely ancestral mechanism, the PRDM9-directed recombination is apparently a derived character with limited phylogenetic distribution [53, 54]. Nevertheless, a common feature of the localization of meiotic recombination events across species seems to be the influence of chromatin structure.

Birds have high CO rates compared to the mammalian sister lineage and also show high within-genome variation in the rate of COs [22, 5457]. The former owes to the fact that avian karyotypes are characterized by a large number of chromosomes and that there is a positive correlation between the amount of COs and the number of chromosomes across organisms [58]). The latter is a consequence of significant variation in chromosome size with numerous small microchromosomes (<5–10 Mb) in which one obligate CO event per chromosome [59] implies high rates of COs per physical unit DNA. However, not much is known about rate heterogeneity at a local scale and what determines the genomic location of CO as well as NCO events in this vertebrate lineage [54]. Addressing these issues is particularly warranted by the fact that birds lack Prdm9 [60], raising the question if the regulation of recombination is more similar to what might be the ancestral mechanism, found in yeast and plants, than to the mechanism found in the mammalian sister lineage [54].

To gain increased insight into recombination in an avian system we localized and characterized recombination events with high resolution by whole-genome re-sequencing of a three-generation pedigree of the collared flycatcher, Ficedula albicollis. We thereby benefitted from the access to a genome assembly with high sequence continuity and with scaffolds anchored, ordered and oriented on chromosomes [55, 61]. Moreover, the availability of a high-density linkage map in this species [55] provides valuable background information on regional CO rate variation across the genome. We identified 325 CO events (at a median resolution of 1.4 kb and with 86% of the events localized to regions < 10 kb), as well as 267 NCO gene conversion events, and used these data to analyse the characteristics and consequences of recombination in an avian system. Our main conclusions from this work are that there is a concentration of recombination events to certain hot-spot regions, which show an association with genes, especially promotor regions and CpG islands. We further find that CO rates are 52% higher in male (3.56 cM/Mb) than in female meiosis (2.28 cM/Mb), that the male CO rate is higher towards chromosome ends, and that there is positive CO interference up to a distance of 14 Mb. The location of NCO events is associated with the location of CO events, while no significant difference between sexes can be observed. Moreover, we find a significant transmission distortion in favour of G and C alleles over A and T alleles at NCOs, providing direct evidence for GC-biased gene conversion in an avian species.


We performed whole-genome re-sequencing (mean autosomal coverage = 42X, range 36.9–45.4X; S1 Table) of 11 collared flycatchers from a three-generation pedigree (Fig 1A) in which 4.434 million segregating SNPs originating from the four grandparents and being informative for phasing (Fig 1B) were identified. A total of 325 meiotic CO events (50–67 per offspring, positions given in S2 Table) were identified in the transmission of gametes from the two F1 parents to the five F2 offspring by mapping the transitions between haploblocks along chromosomes (Fig 1C). Due to a high degree of nucleotide diversity (π) in the population (mean π = 3.6 x 10−3; [9, 61]) and that deep sequencing allowed SNPs to be called at a high rate, the position of CO events could be identified with high accuracy (S1 Fig). The median interval between recombinant SNP markers was 1,513 bp, or 1,360 bp if only considering events in genomic regions without assembly gaps. Eighty per cent of all CO events could be mapped with a resolution of <5 kb, and 86% <10 kb, with similar resolution in all five F2 offspring (S3 Table). After very stringent filtering (see Methods) we further identified a total of 267 NCO gene conversions spread across the flycatcher genome (S4 Table). Given the stringent filtering and that the power to detect NCOs is low, the set of identified NCOs likely represents only a subset of all such events.

Fig 1. The flycatcher pedigree and illustration of crossover detection.

(a) The three-generation pedigree used in this study. (b) Schematic illustration of phasing. A SNP can be phased when the grandparental genotypes differ from each other (either because they are homozygous for different alleles, or one is homozygous and the other is heterozygous) and the F1 is heterozygous. If the phase can be traced also in the F2 generation, it is possible to pinpoint recombination events in the F1 gametes. (c) Haploblocks identified in the five largest chromosomes in one male F2 offspring. The left chromosome in each pair represents the paternally transmitted chromosome and the right the maternally transmitted chromosome. Light blue is the contribution from the paternal grandfather, green the paternal grandmother, orange the maternal grandfather, and red the maternal grandmother. Note that the Z chromosome does not recombine in female meiosis, with the exception of in the pseudoautosomal region (PAR, insert).

The number of CO events per chromosome and meiosis ranged between 0–6. The amount of COs per chromosome as reflected in number of CO events was in excellent agreement with predictions from genetic distances observed in linkage analysis (Pearson’s r = 0.95, Table 1), providing strong support for the overall accuracy of the detection of CO events. Moreover, regional (200 kb windows) CO rate estimates based on linkage analysis are available for the collared flycatcher genome [55] and windows corresponding to the location of CO events detected in the present study had a significantly higher linkage-based CO rate (6.27 cM/Mb) than windows without CO events (3.65 cM/Mb; t-test, p = 6.3 x 10−7). The total amount of observed COs per meiosis corresponded to a sex-averaged autosomal genetic distance of 3,030 cM, very close to that obtained in linkage analysis (3,067 cM; [55]). The average CO rate in autosomes (data available for 30 autosomes) was 3.08 cM/Mb but the rate was highly variable among chromosomes and showed a strong non-linear correlation with chromosome size (Fig 2). For chromosomes >50 Mb, the mean CO rate was 1.98 cM/Mb and for chromosomes <10 Mb the mean rate was 13.02 cM/Mb.

Fig 2. The relationship between chromosome size and crossover rate.

The figure shows the sex-average crossover rate per chromosome.

Table 1. Recombination distance per chromosome calculated from the number of observed recombination events in the pedigree and linkage map length from the corresponding chromosomes.

Linkage map data are from Kawakami et al. [55] and refer to the sex-average best-order map length per chromosome.

Of 305 detected autosomal CO events, 119 were of maternal origin and 186 of paternal origin. This corresponds to total map distances of 2,380 cM in female meiosis (2.28 cM/Mb) and 3,720 cM (3.56 cM/Mb) in male meiosis, i.e. 56% higher CO rate in males than in males. Out of a total of 20 events on the Z chromosome, two were of maternal origin and were located in the ≈ 0.6 Mb pseudoautosomal region [62]. This confirms a very high rate (67 cM/Mb, similar to what has been observed in linkage analysis; [62]) of COs in this short region, which corresponds to ≈1% of the Z chromosome and which is the only region where the Z chromosome and W chromosome pairs in female meiosis. Contrary to the pattern observed for COs, the average number of NCOs that occurred during male meiosis (25.4) was not statistically different to the number of NCOs that occurred during female meiosis (27.4; t-test, p = 0.506). Only a single NCO event was identified on the Z chromosome and was of paternal origin.

It is often assumed that one obligate CO per chromosome is necessary for proper segregation during meiosis, irrespective of chromosome size [63] (though there are organisms in which this does not apply, like the absence of recombination in male meiosis of Drosophila). We observed many instances of transmitted chromosomes without a detected CO event. This is not surprising given that 50% of the gametes from CO events will be non-recombinants. Thus, when there is close to only one CO event per chromosome on average, there should be about as many gametes with an observable CO event as without. To corroborate this assumption we focused on maternal transmission of the smallest microchromosomes (as the overall rate of COs was lower in females and the likelihood for more than one CO event in chromosomes < 10 Mb (n = 9) can be considered low, and was not observed). As predicted, the number of instances with one CO event (24) was similar to the number of instances without a detected CO event (21) in these small chromosomes. The distribution of transmission of recombinant versus non-recombinant chromosomes for the whole data set is shown in S2 Fig The higher number of transmissions of non-recombinant chromosomes in female than in male meiosis is a logical consequence of the lower rate of COs in females.

There was a general trend of a higher frequency of CO events towards the ends of chromosomes (frequency measured in 10 Mb windows; Fig 3A, S3 Fig), primarily in chromosomes 50–100 Mb in size and in male meiosis. Moreover, within the terminal 10 Mb there was a markedly higher frequency of events towards the ends (Fig 3B; frequency measured in 1 Mb windows). Interestingly, also at this scale there was distinct difference between the sexes in that the frequency of CO events in male meiosis increased towards the very end while there was no CO event observed in the terminal 1 Mb in female meiosis (Mann-Whitney U-test, z = 1.87, p = 0.030). A similar trend was seen between the frequency of NCOs and distance to chromosome end (S4 Fig).

Fig 3. Distribution of the number of crossover events in relation to distance to nearest chromosome end.

(a) chromosomes >100 Mb (males, black; female, white) and chromosomes 50–100 Mb (males, grey; females dotted) for 10 Mb intervals, and (b) the terminal 10 Mb of all chromosomes in 1 Mb intervals (males, black; female, white).

We tested whether the location of one CO event affected the location of other events on the same chromosome or if the locations were independent of each other. There was evidence for positive interference–lowered likelihood of two nearby CO events–up to a distance of 14 Mb (Fig 4). Interestingly, out of 107 detected double CO events, only 28 were detected in female meiosis, while 79 were detected in male meiosis. Together with positive interference, this could explain why the observed increase in CO rate towards chromosomes ends was more pronounced in males than in females.

Fig 4. Mean crossover interference in relation to the distance of double crossovers.

Interference was measured as the coefficient of coincidence (CoC). The horizontal dashed line at a CoC of 1 indicates an interference of 0.

The distribution of CO events along chromosomes in the flycatcher genome indicates that there are several distinct regions with a concentration of CO events, i.e. CO hot-spots (S5S7 Figs). We identified 19 regions on 12 different chromosomes with two or more CO events from independent meiosis localized to less than 100 kb apart, demonstrating a highly skewed distribution of CO events. Randomly placing CO events along chromosomes indicated that the likelihood for this to happen by chance was <0.001. Excluding CO events that overlapped with gaps between scaffolds, randomization revealed that the likelihood for the observed amount of co-localized CO events by chance was <0.005.

There was a concentration of CO events to genic regions (Fig 5A). To analyse the association between genes and COs in some further detail we divided the genome into promoter regions (2 kb upstream of transcription start site, TSS), first exons, first introns, other exons, other introns and intergenic DNA. The CO rate was highest in promoter regions (1.85 times the intergenic rate), followed by first exons and other exons (Fig 5B). Among assembled parts of the genome, the number of CO events per bp in promoter regions was significantly higher than in intergenic DNA (Fisher’s Exact test, p = 0.018). No statistically significant differences were detected between other comparisons, which may be due to the limited number of CO events in exons and introns. Moreover, there was a significant association between COs and CpG islands (p = 8.72x10-9). Given that CpG islands are prevalent upstream of genes, the overrepresentation of CO events in promoters and CpG islands is likely not independent from each other. Besides, the GC content of CO regions (45.6%) was higher than in the genomic background (41.9%). This may in part be due to the fact that CpG islands are high in GC content and to the disproportionate number of CO events on the small microchromosomes (given one obligate CO, independent of chromosome size), and the fact that GC content increases with decreasing chromosome size in birds [22]. However, the GC content of CO regions on each chromosome was significantly higher than the genomic background on the respective chromosomes (paired t-test, p = 0.0014). We did not find any evidence for a higher repeat density in CO regions (9.1% vs. 8.2%; p = 0.65).

Fig 5. The association between recombination and genes.

(a) Distribution of recombination events in relation to distance to the closest gene. The large bar at position 0 represents the number of events overlapping with genes. (b) Density of CO events in different sequence categories of genes, and in intergenic DNA. c) Density of NCO events in different sequence categories of genes, and in intergenic DNA.

The small sample size of NCO events gives limited power for statistical analyses of the relative abundance of NCO events in different functional categories (Fig 5C). However, a higher density in first exons, i.e. close to TSS, compared to intergenic regions was close to significant (odds ratio = 2.495, p = 0.081), and would resemble the situation for CO events. A significant overlap in the localization of CO and NCO events compared to random expectations was observed (odds ratio = 6.321, p = 0.0041), providing further support for a common mechanism of regulation.

The genomic landscape of species divergence in flycatchers is characterized by the presence of numerous (≈50) ‘differentiation islands’ spread across the genome, evident as distinct FST peaks in comparisons between species [9, 61]. These islands cover roughly 7% of the genome and may primarily result from lowered Ne due to linked selection in regions of low CO rate [9]. If the genomic locations of CO and NCO events and the 50 differentiation islands were unrelated, we would expect to see approximately 7% of these events to overlap with islands by chance. However, only 11 out of 325 CO events overlapped with islands (odds ratio = 0.454, p = 0.0067). In contrast, NCO events were over-represented in islands; 33 out of 266 NCO events overlapped (odds ratio = 1.837, p = 0.0026). Considering both types of recombination events taken together, their distribution relative to differentiation islands did not differ significantly from 7% of overlap.

Both CO and NCO can lead to tracts of gene conversion close to the location of DSBs [12, 64]. However, since we only trace one product of meiosis, we cannot track gene conversion tracts at CO events. Of the 267 NCO events, 229 involved sites segregating for one ‘weak’ (‘W’; A or T) and one ‘strong’ (‘S’; G or C) allele. We then counted the number of times a weak allele was converted by a strong allele (W>S) and the number of times a strong allele was converted by a weak allele (S>W). If transmission of alleles upon gene conversion is a random process there should be about as many events of one category as of the other. However, there was a significant excess of W>S conversions (Binomial test, p = 0.012), with a biased transmission of 59% (95% CI = 0.52–0.65). This corresponds to a transmission distortion (c) of 0.18 and provides direct evidence for GC-biased gene conversion in an avian species.


Whole-genome re-sequencing provides high accuracy in mapping the localization of recombination events, with resolution in the present type of study basically determined by the density of segregating (and informative) sites in the pedigree and assuming that sequencing depth is sufficient to accurately call most variants in the analysed individuals. In our study, the median resolution of CO events was 1.4 kb and 86% of all 325 events could be localized to intervals < 10 kb. This represents an improved resolution of the localization of CO events by several orders of magnitude compared to data from even the densest linkage maps of birds (e.g. [24, 5557, 65]). Moreover, compared to recombination studies in the mammalian sister lineage, it also implies higher resolution than in a similar pedigree-sequencing study of chimpanzee (median interval of detected CO events of 7 kb; [32]) and in genome-wide sperm-sequencing studies of humans (13–45% of CO events mapped within <30 kb; [34, 66]) a likely consequence of the higher density of polymorphic sites in flycatchers than in primates.

In comparison to fine-scale CO rate estimates based on the extent of linkage disequilibrium (LD) inferred from whole-genome re-sequencing of population samples [67], pedigree-sequencing cannot realistically reach the same genome-wide coverage in rate estimation since patterns of LD reflect the landscape of a very large number of historically accumulated CO events across the whole genome. However, LD is not only affected by the rate of COs but also by demography and selection. On the other hand, pedigree-sequencing directly pinpoints the occurrence of CO as well as NCO events and therefore provides an instantaneous picture of current recombination patterns. One limitation of our study is that we focused on a single three-generation family and a general caveat is thus that the results may depend on the particular genetic background of the four individuals of the P generation and recombination characteristics of the two F1 parents. More extended pedigree-sequencing could be used to detect variation in rates and patterns of recombination between individuals, sexes and populations.

With a total genetic distance of just above 3,000 cM, the overall rate of COs in collared flycatcher is similar to that reported in chicken [24]. As judged from total map lengths in linkage analysis, the rate is higher than in two bird species that are more closely related to flycatcher than chicken, namely great tit (Parus major, ≈1,900 cM, [56]) and zebra finch (Taeniopygia guttata, 1,100–1,500 cM, [57, 65]). Some difficulty in comparing map lengths in birds follows from the disproportionate localization of COs on the many microchromosomes. These are often poorly covered, or even uncovered, by genetic markers in linkage analysis, making estimates of the total map distance sensitive to marker abundance and distribution. We suggest that this explains part of the differences in overall CO rate seen among avian species. However, even after taking these aspects into account, biologically meaningful differences probably remain.

Recombination hot-spots and their location relative to genes

Studies in several species of plants and fungi have shown that recombination events are concentrated in hot-spots in close proximity to TSSs [28, 3943]. This may be a consequence of the common pattern that both transcription and recombination are facilitated in open chromatin [68, 69]. In humans and mouse, PRDM9 directs recombination away from TSS [51, 52, 67, 70, 71]. The absence of an active Prdm9 gene in avian genomes [60] prompts the hypothesis that recombination in this vertebrate lineage resembles the ancestral mechanism of regulation and is associated with proximity to TSSs, similar to plants and fungi. Our observations confirmed this hypothesis: the highest rate of CO events was seen in promoter regions and first exons, with statistical support for the rate in the former being higher than in intergenic regions. We also found an association between COs and CpG islands, which commonly function as promoters by destabilizing nucleosomes and attracting proteins that create a transcriptionally permissive chromatin state [72]. Also NCO events tended to be most common close to TSSs and, overall, there was evidence for an overlap in the distribution of CO and NCO events. As pointed out by Lichten, this suggests “that the picture in mammals may be the exception rather than the rule” [73].

We note that there are exceptions to an increased rate of recombination close to TSS in the absence of Prdm9. In both Drosophila melanogaster [74] and D. pseudoobscura [75], recombination is reduced around TSS. Furthermore, fruit flies [33, 75], worms [76] and honeybee [77] have recombination landscapes that are relatively homogenous without distinct hotspots. Apparently, despite being such a widespread phenomenon across the tree of life, recombination has evolved distinct characteristics in different lineages.

The usage of human and mouse recombination hot-spots has a high turnover rate [78, 79]. This owes in part to large allelic variation of, and positive selection on, the DNA-binding residues of zinc fingers [53], such that it represents an allelic turnover of the binding protein. However, erosion of binding motifs in target sequence due to recombination-induced mutation or gene conversion probably also occurs [2, 47]. The concentration of recombination events to hot-spot regions in the absence of an active Prdm9 gene, like what we found in this study of flycatchers, could potentially mean that the hot-spot landscape of recombination in such lineages remains relative stable (cf. [80]). Evidence for this was recently found in an avian study of finch species [54] and in analyses of divergent Saccharomyces species [81], where the location of hot-spots was found to be conserved over considerable evolutionary time.

Heterochiasmy and the overall distribution of recombination along chromosomes

We found a higher rate of COs in male than in female meiosis, i.e. higher rate in the homogametic than in the heterogametic sex. Since we only measured the rate of COs in one individual of each sex, we cannot formally exclude the possibility that part of the observed rate difference was due to different genetic backgrounds rather than to sex per se. However, the higher male CO rate accords with collared flycatcher linkage map data [55, 82]. Birds do not uniformly follow the Haldane-Huxley rule (reduced CO rate in the heterogametic sex, [83]) since there in addition to species with male-biased CO rate [84] are others with female-biased CO rate [85] or similar rates of COs in the two sexes [24, 56, 65, 86]. Haldane [87] and Huxley [88] suggested that reduced CO rate in the heterogametic sex was a pleiotropic consequence of selection against CO between diverging X and Y (or Z and W) sex chromosomes. Several alternative hypotheses have subsequently been put forward (see [85]), of which some could potentially explain variation in the relative rates of male and female COs within organism groups. For example, sexual selection may select for reduced CO rate (to maintain favourable allelic combinations) in the sex with the largest variance in reproductive success [89], potentially setting the stage for a relationship between the intensity of sexual selection (or sexual antagonism) and male-to-female CO rate ratio [90, 91].

As for other animal groups (e.g. [92]), linkage analysis in several bird species has revealed a general trend of increased rates of COs towards chromosome ends [24, 55, 56], the extent of which apparently varies among species with the most pronounced end effect so far seen in zebra finch [57, 65]. Our observations add to and complement this picture by demonstrating a sex difference in the distribution of COs along chromosomes. Specifically, the increased CO rate towards chromosome ends was mainly seen in males, similar to the situation in humans and mice [93], while female CO events were more evenly distributed. The findings of positive CO interference up to about 14 Mb and a higher incidence of double COs in males than in females are compatible with increased rates of COs towards chromosome ends in males. Only a limited number of recombination-initiating DSBs eventually result in COs, whose spatial distribution is tightly regulated through the process of CO interference that reduces the possibility of two nearby CO events [94]. As a result, double COs tend to be directed towards the respective ends of chromosomes. Measures of the extent of interference vary among taxa with a tendency for shorter distances seen in organisms with higher CO rates than in organisms with lower CO rates [95]. Estimates of CO interference for human and mice, for example, range between 20 and 140 Mb [35, 96]. The relevant metric in terms of CO interference is though the physical distance in μm along the bivalent, not the “genomic” distance in Mb along DNA. The distance over which interference occur thus depends on the degree of compaction of chromosomes at the leptotene stage of meiosis. This may explain sex-differences in the spatial distribution of COs, since males and females show varying degrees of compaction of the chromosomes [96]. Detailed mapping of recombination events in human sperm and oocytes has documented that CO interference is more pronounced in males than in females [34, 35]. It may very well be that other meiotic characteristics of spermatogenesis and oogenesis contribute to sex-differences in patterns of recombination, like differences in the time allotted to the bouquet formation at telomeres [93].

An interesting observation with respect to the distribution of COs along chromosomes was the significant under-representation of CO events in genomic regions defined as differentiation islands. This accords with the findings that differentiation islands are concentrated, if not limited, to regions of the genome corresponding to CO desserts, as judged by CO rate data from linkage maps, and that there is an overall positive correlation between CO rate and FST [9]. As such, this corroborates the notion that CO rate drives the genomic landscape of species differentiation in Ficedula flycatchers [9]. The rationale for this inference is that the prevalence of the diversity-reducing (Ne-reducing) effects of linked selection increases with decreasing CO rate [97]. Because the role of genetic drift on differentiation in turn increases with decreasing Ne, this means that variation in the degree of differentiation across the genome is compatible with a neutral model, without the need to invoke selection or varying degree of gene flow. Moreover, since both types of recombination events taken together did not differ significantly from a random overlap and NCO events were over-represented in islands, this indicates that DSBs that occur in these differentiation islands are preferentially assigned as NCOs.

GC-biased gene conversion (gBGC)

We found a significant transmission distortion at NCO events with the ‘strong’ (G, C) allele transmitted at 59% of all events involving one ‘strong’ and one ‘weak’ (A, T) allele (transmission distortion, c = 0.18). Detection of a transmission distortion for strong alleles has so far been limited to studies of humans and yeast [12, 14, 64, 98, 99], and is likely explained by GC-biased gene conversion (gBGC). gBGC is a process associated with meiotic recombination, which favours strong base-pairs over weak base-pairs at weak:strong mismatches in heteroduplex DNA formed as part of the repair mechanism of DSBs. This ultimately leads to a preferential transmission of strong alleles over weak alleles close to recombination-initiating DSBs. As a consequence, the local rate of recombination is expected to show a positive correlation with the local GC content. Indeed, such indirect evidence for gBGC has been observed across a wide range of taxa [11], while direct detection is much more rare.

In yeast a transmission distortion of 0.057 was observed at CO events, but no significant distortion was observed at NCO events [12]. In humans there is evidence for a transmission distortion associated with both CO and NCO events [64, 98]. A recent genome-wide study of NCO events in humans estimated a transmission distortion of 0.36 [99]. The estimate that we report here for flycatchers falls below the estimate for humans. However, the net impact of gBGC on the evolution of base composition does not only depend on the strength of the transmission distortion, but also on the number of weak:strong mismatches in heteroduplex DNA, the recombination rate and Ne [11]. Given higher SNP density [9], higher recombination rate [55] and larger Ne in flycatchers [100] compared to humans, this might readily account for a higher genome-wide GC content in flycatchers compared to humans.

We have previously suggested that the slow rate of karyotypic evolution in birds will promote a conserved genomic landscape of recombination rate variation and thereby facilitate the evolutionary build-up of genomic signatures of recombination, like the effect of gBGC on base composition [101]. The presence of recombination hot-spots coupled with a stable hot-spot landscape in the absence of Prdm9 further accentuates the influence of recombination rate variation on avian molecular evolution. For example, in a recent flycatcher study we found an at first glance unexpected absence of a correlation between recombination rate and the rate of non-synonymous substitutions [102]. However, the patterns changed when GC-biased gene conversion was taken into account and weak-to-strong and strong-to-weak substitutions were separately analysed.

Materials and Methods

Ethics statement

This study was approved by Linköpings djuretiska nämnd, Linköpings tingsrätt, Sweden (Dnr 21–11).

Data generation

We sequenced a three-generation pedigree (Fig 1A) of 11 collared flycatchers, sampled in the field from a natural population on the Baltic Sea island Öland (Sweden). The four birds in the P generation showed no evidence of being closely related. Sequencing was done to approximately 40X coverage on an Illumina HiSeq 2000 instrument with paired-end reads of 100 bp and libraries (insert size of 450 bp) constructed using the TruSeq Nano sample preparation kit (Illumina) (European Nucleotide Archive PRJEB12616). DNA were prepared from blood samples stored in 96% ethanol using a standard proteinase K digestion/phenol-chloroform purification protocol. The reads were aligned to the collared flycatcher reference genome FicAlb1.5 (GenBank Accession GCA_000247815.2) with bwa 0.7.5a [103] and de-duplicated, recalibrated and cleaned with GATK 3.2.2 [104, 105].

Variant calling and de novo discovery

Single nucleotide polymorphisms (SNPs) were called with GATK's HaplotypeCaller and GenotypeGVCFs (version 3.3.0). Variant Quality Score Recalibration (VQSR) was performed according to the GATK's Best Practice [106], using known SNP positions from genotyping with a 50k SNP-chip [107] and the 20% top scoring sites for training. Sites that failed the 99.9% tranche threshold were removed. We also removed sites in repetitive regions using a combination of RepeatMasker v3.2.9 (Smit, AFA, Hubley, R & Green, P. <>) and a flycatcher-specific repeat library [108], Tandem Repeats Finder v4.07 [109] and an in-house perl script masking homopolymers longer than 10bp that were not already masked. To further reduce the number of mis-genotyped sites, we applied a coverage filter removing sites where any of the involved individuals were covered by less than 15 reads, or covered by more than twice the average autosomal coverage; the latter criterion was applied to reduce the risk of using collapsed regions where sites potentially could be mis-called as heterozygous due to differences between sequence copies. We also filtered sites with low genotype quality (GQ<30), which corresponds to that the probability of choosing the wrong genotype is less than 0.001. Sites violating Mendelian inheritance or with more than two called alleles were removed. All these measures ensured stringent filtering.

Handling of the Z chromosome

The entire genome including the Z chromosome was variant-called as diploid. To avoid incorrectly assigned genotypes for females (which in birds represent the heterogametic sex, with one copy of the Z chromosome plus the female-specific W chromosome), the Z chromosome was called separately as haploid (flag -ploidy 1 in GATK) in the four females (two in the P generation and one each in the F1 and F2 generations); the pseudo-autosomal region of the Z chromosome [62] was excluded from this analysis. VQSR could not be used for this set due to lack of a proper training set, but we filtered using coverage and genotype quality. 15X was used as the lower coverage threshold as above, and maximum coverage was set to the mean autosomal coverage (i.e., twice the expected coverage for the Z chromosome). The new haploid calls for the females where then added to the data set.

Extracting informative sites

With a three-generation family it is possible to phase the F1 generation into chromosome-level haplotypes and thereby detect recombination events in the transmission of gametes to the F2 generation (Fig 1C). For a variant site to be informative in phasing, it is required that an F1 individual is heterozygous and that its two parents have different genotypes; this situation makes it possible to trace each F1 allele to one of the parents in the P generation. It also requires that F1 alleles are traceable in the F2 generation and the F1 individual's partner and offspring must therefore not both be heterozygous (S8 Fig). Phased sites were grouped into haploblocks in the five F2 individuals. Similar to [32] we assumed that there would be at most one recombination event per Mb interval and considered shorter blocks as the likely result of phasing or genotyping error (see further below).

Identification of CO events

CO events were localized to the genomic interval between the outermost SNPs of two adjacent haploblocks in the F2 individuals. The resolution of intervals varied depending on the density of informative SNPs in the regions in question. We limited downstream analyses to events that were localized with a resolution <5 kb. The location of 25 recombination intervals overlapped with gaps between scaffolds meaning that the precise size could not be determined.

Identification of NCO events

NCO events are suggested at informative positions with a phase that does not match the surrounding block. However, this will also be the case whenever there is an incorrect genotype call in any of the involved individuals. Manual inspection in IGV [110] of a subset of phase-mismatched sites showed that most of them had visible problems such as high coverage (close to our upper threshold), unequal numbers of reads supporting the two alleles at a site, reads supporting more than two visible alleles, clusters of nearby polymorphic sites, and overlaps with insertions and deletions. There was also an excess of sites at which all females were heterozygous and all males were homozygous for the reference allele, indicative of reads from W-linked sequences mapping to autosomal or Z-linked loci in the male-derived genome assembly.

To reduce the above-mentioned problems in inferring potential NCO events we applied several additional stringent filtering criteria to the set of informative sites selected for the identification of CO events. First we used a more strict VQSR tranche threshold of 90%. We further excluded all sites overlapping with, or present within 10 bp of, indels called by GATK and clusters of SNPs that had more than three called SNPs within 30bp in the full VQSR-filtered file (containing all individuals). We also excluded larger clusters of SNPs with frequent and alternating haplotype shifts; based on visual inspecting we set the threshold to no more than two deviating sites in 5 kb. Next, we removed sites that had reads supporting more than two alleles, sites where one of the alleles was supported by < 25% of the reads, and sites where all females were heterozygous and all males were homozygous for the reference allele. We considered this filtering necessary to remove ambiguous sites although it may have come to the price of underestimating the occurrence of NCO events. Because of this we do not investigate the relative frequencies of CO and NCO events.

Associations between recombination events and genomic regions

Collared flycatcher genes were downloaded from Ensembl (release 73, assembly version FicAlb_1.4) and translated to the latest assembly version with chromosomes using an in-house script. For calculating the distance between CO events and genes, we restricted the analysis to CO events with a resolution <5 kb. Depending on the orientation of the closest gene we assigned the CO event to the upstream or the downstream flank. If the CO event overlapped with a gene, we assigned it a distance of 0. In a second step we assigned CO events into six different classes of genomic regions; intergenic, promoter (defined as 2 kb upstream of the transcription start site, TSS), first exon, first intron, other exons and other introns. A CO event that overlapped several classes was assigned values to each of these classes proportional to the length of the overlap. We repeated the analysis for NCO events. Next, CpG islands (CGIs) were identified for the hard-masked flycatcher genome using CpGcluster (version 1.0) with default parameter settings [111]. In order to assess the association between CGIs and CO events, the number of overlapping CGIs was compared to the genome-wide average. All statistics were calculated and plotted with R version 3.0.2 (

CO interference

We used the coefficient of coincidence (CoC) to assess the strength of CO interference [28]. CoC was computed as the number of observed over expected double COs counted in a 1Mb sliding windows approach. This provided us with a sex-average and genome-average CoC.

Supporting Information

S1 Table. Mean autosomal sequence coverage for each of the 11 birds included in the study.


S2 Table. Positions of the recorded CO events.


S3 Table. Summary statistics for recombination events detected in F2 offspring.


S4 Table. Positions of recorded NCO events.


S1 Fig. Distribution of the length (kb) of identified CO intervals.

Only regions shorter than 20 kb are shown; there are 24 additional intervals larger than 20 kb.


S2 Fig. The distribution of transmission of recombinant versus non-recombinant chromosomes for the whole data set, separate for females and males.


S3 Fig. Relationship between genetic distance and physical length.

Cumulative, sex-average genetic distance along chromosomes obtained from the distribution of CO events are shown for the six largest chromosomes.


S4 Fig. Distribution of the number of non-crossover events in relation to distance to nearest chromosome end.

(a) chromosomes >100 Mb (males, black; female, white) and chromosomes 50–100 Mb (males, grey; females dotted) for 10 Mb intervals, and (b) the terminal 10 Mb of all chromosomes in 1 Mb intervals (males, black; female, white).


S5 Fig. Genomic distribution of CO events on flycatcher chromosomes 1–4.

Maternal events are shown to the left of each chromosomes in red and paternal events are shown to the right in blue. Each symbol (diamond, triangle, box, cross, plus) represents one of the five F2 offspring in which CO events were observed.


S6 Fig. Genomic distribution of CO events on flycatcher chromosomes 4A-12, and Z.

Maternal events are shown to the left of each chromosomes in red and paternal events are shown to the right in blue. Each symbol (diamond, triangle, box, cross, plus) represents one of the five F2 offspring in which CO events were observed.


S7 Fig. Genomic distribution of CO events on flycatcher chromosomes 13–28.

Maternal events are shown to the left of each chromosomes in red and paternal events are shown to the right in blue. Each symbol (diamond, triangle, box, cross, plus) represents one of the five F2 offspring in which CO events were observed.


S8 Fig. Details of phasing.

Phasing of haplotypes originating from the paternal (left) and maternal (right) grandparents. Genotypes are shown only for the five individuals that need to be considered for the line in question, and are given as 0 (reference allele) and 1 (alternative allele). In the paternal example, the first and third SNPs in the F2 are informative while the second SNP is uninformative because both 0 and 1 can come from either F1 parent. In the maternal example, only the third SNP in the F2 can be traced back to the P generation and is informative. The first SNP can be traced back to the F1 but is not informative.



Sequencing and initial bioinformatic analysis were performed by the SNP&SEQ Technology Platform in Uppsala. The facility is part of the National Genomics Infrastructure (NGI) Sweden and Science for Life Laboratory. The SNP&SEQ Platform is also supported by the Swedish Research Council and the Knut and Alice Wallenberg Foundation. Computations were performed on resources provided by the Swedish National Infrastructure for Computing (SNIC) through Uppsala Multidisciplinary Center for Advanced Computational Science (UPPMAX). The authors acknowledge helpful comments made by three anonymous reviewers.

Author Contributions

Conceived and designed the experiments: HE LS. Performed the experiments: LS. Analyzed the data: LS CFM HE. Contributed reagents/materials/analysis tools: AQ. Wrote the paper: HE LS CFM.


  1. 1. Keeney S, Giroux CN, Kleckner N. Meiosis-Specific DNA Double-Strand Breaks Are Catalyzed by Spo11, a Member of a Widely Conserved Protein Family. Cell. 1997;88(3):375–84. pmid:9039264
  2. 2. Webster MT, Hurst LD. Direct and indirect consequences of meiotic recombination: implications for genome evolution. Trends in Genetics. 2012;28(3):101–9. pmid:22154475
  3. 3. Hill WG, Robertson A. The effect of linkage on limits to artificial selection. Genetical Research. 1966;8:269–94. pmid:5980116
  4. 4. Charlesworth B, Charlesworth D. The degeneration of Y chromosomes. Philosophical Transactions of the Royal Society of London B: Biological Sciences. 2000;355(1403):1563–72. pmid:11127901
  5. 5. Lercher MJ, Hurst LD. Human SNP variability and mutation rate are higher in regions of high recombination. Trends in Genetics. 2002;18:337–40. pmid:12127766
  6. 6. Charlesworth B, Campos JL. The relations between recombination rate and patterns of molecular variation and evolution in Drosophila. Annual Review of Genetics. 2014;48(1):383–403.
  7. 7. Kulathinal RJ, Bennett SM, Fitzpatrick CL, Noor MAF. Fine-scale mapping of recombination rate in Drosophila refines its correlation to diversity and divergence. Proceedings of the National Academy of Sciences USA. 2008;105(29):10051–6.
  8. 8. Spencer CCA, Deloukas P, Hunt S, Mullikin J, Myers S, Silverman B, et al. The influence of recombination on human genetic diversity. PLoS Genetics. 2006;2(9):e148. pmid:17044736
  9. 9. Burri R, Nater A, Kawakami T, Mugal CF, Olason PI, Smeds L, et al. Linked selection and recombination rate variation drive the evolution of the genomic landscape of differentiation across the speciation continuum of Ficedula flycatchers. Genome Research. 2015;25:1656–65. pmid:26355005
  10. 10. Hill WG, Robertson A. Linkage disequilibrium in finite populations. Theoretical and Applied Genetics. 1968;38:226–31. pmid:24442307
  11. 11. Mugal CF, Weber CC, Ellegren H. GC-biased gene conversion links the recombination landscape and demography to genomic base composition. BioEssays. 2015.
  12. 12. Lesecque Y, Mouchiroud D, Duret L. GC-biased gene conversion in yeast Is specifically associated with crossovers: molecular mechanisms and evolutionary significance. Molecular Biology and Evolution. 2013;30(6):1409–19. pmid:23505044
  13. 13. Duret L, Galtier N. Biased gene conversion and the evolution of mammalian genomic landscapes. Annual Review of Genomics and Human Genetics. 2009;10(1):285–311.
  14. 14. Mancera E, Bourgon R, Brozzi A, Huber W, Steinmetz LM. High-resolution mapping of meiotic crossovers and non-crossovers in yeast. Nature. 2008;454(7203):479–85. pmid:18615017
  15. 15. Coop G, Wen X, Ober C, Pritchard JK, Przeworski M. High-resolution mapping of crossovers reveals extensive variation in fine-scale recombination patterns among humans. Science. 2008;319(5868):1395–8. pmid:18239090
  16. 16. Ma L, O'Connell JR, VanRaden PM, Shen B, Padhi A, Sun C, et al. Cattle sex-specific recombination and genetic control from a large pedigree analysis. PLoS Genetics. 2015;11(11):e1005387. pmid:26540184
  17. 17. McGaugh SE, Heil CSS, Manzano-Winkler B, Loewe L, Goldstein S, Himmel TL, et al. Recombination modulates how selection affects linked sites in Drosophila. PLoS Biology. 2012;10(11).
  18. 18. Jiang H, Li N, Gopalan V, Zilversmit M, Varma S, Nagarajan V, et al. High recombination rates and hotspots in a Plasmodium falciparum genetic cross. Genome Biology. 2011;12(4):R33. pmid:21463505
  19. 19. Smukowski CS, Noor MA. Recombination rate variation in closely related species. Heredity. 2011;107:496–508. pmid:21673743
  20. 20. Wilfert L, Gadau J, Schmid-Hempel P. Variation in genomic recombination rates among animal taxa and the case of social insects. Heredity. 2007;98:189–97. pmid:17389895
  21. 21. Dumont BL, Payseur BA. Evolution of the genomic rate of recombinaton in mammals. Evolution. 2008;62(2):276–94. pmid:18067567
  22. 22. ICGSC. Sequence and comparative analysis of the chicken genome provide unique perspectives on vertebrate evolution. Nature. 2004;432(7018):695–716. pmid:15592404
  23. 23. Lenormand T. The evolution of sex dimorphism in recombination. Genetics. 2003;163(2):811–22. pmid:12618416
  24. 24. Groenen MAM, Wahlberg P, Foglio M, Cheng HH, Megens H-J, Crooijmans RPMA, et al. A high-density SNP-based linkage map of the chicken genome reveals sequence features correlated with recombination rate. Genome Research. 2009;19(3):510–9. pmid:19088305
  25. 25. Slavov GT, DiFazio SP, Martin J, Schackwitz W, Muchero W, Rodgers-Melnick E, et al. Genome resequencing reveals multiscale geographic structure and extensive linkage disequilibrium in the forest tree Populus trichocarpa. New Phytologist. 2012;196(3):713–25. pmid:22861491
  26. 26. Baudat F, Imai Y, de Massy B. Meiotic recombination in mammals: localization and regulation. Nature Reviews Genetics. 2013;14(11):794–806. pmid:24136506
  27. 27. Arnheim N, Calabrese P, Tiemann-Boege I. Mammalian Meiotic Recombination Hot Spots. Annual Review of Genetics. 2007;41(1):369–99.
  28. 28. Li X, Li L, Yan J. Dissecting meiotic recombination based on tetrad analysis by single-microspore sequencing in maize. Nature Communications. 2015;6.
  29. 29. Wijnker E, Velikkakam James G, Ding J, Becker F, Klasen JR, Rawat V, et al. The genomic landscape of meiotic crossovers and gene conversions in Arabidopsis thaliana. eLife. 2013;2:e01426. pmid:24347547
  30. 30. Lu P, Han X, Qi J, Yang J, Wijeratne AJ, Li T, et al. Analysis of Arabidopsis genome-wide variations before and after meiosis and meiotic recombination by resequencing Landsberg erecta and all four products of a single meiosis. Genome Research. 2012;22(3):508–18. pmid:22106370
  31. 31. Yang S, Yuan Y, Wang L, Li J, Wang W, Liu H. Great majority of recombination events in Arabidopsis are gene conversion events. Proceedings of the National Academy Sciences USA. 2012;109:20992–7.
  32. 32. Venn O, Turner I, Mathieson I, de Groot N, Bontrop R, McVean G. Strong male bias drives germline mutation in chimpanzees. Science. 2014;344(6189):1272–5. pmid:24926018
  33. 33. Comeron JM, Ratnappan R, Bailin S. The many landscapes of recombination in Drosophila melanogaster. PLoS Genetics. 2012;8:e1002905. pmid:23071443
  34. 34. Lu S, Zong C, Fan W, Yang M, Li J, Chapman AR, et al. Probing meiotic recombination and aneuploidy of single sperm cells by whole-genome sequencing. Science. 2012;338(6114):1627–30. pmid:23258895
  35. 35. Hou Y, Fan W, Yan LY, Li R, Lian Y, Huang J, et al. Genome analyses of single human oocytes. Cell. 2013;155(7):1492–506. pmid:24360273
  36. 36. Xu S, Ackerman MS, Long HA, Bright L, Spitze K, Ramsdell JS, et al. A male-specific genetic map of the microcrustacean Daphnia pulex based on single-sperm whole-genome sequencing. Genetics. 2015;201(1):31–+. pmid:26116153
  37. 37. Ottolini CS, Newnham LJ, Capalbo A, Natesan SA, Joshi HA, Cimadomo D, et al. Genome-wide maps of recombination and chromosome segregation in human oocytes and embryos show selection for maternal recombination rates. Nature Genetics. 2015;47(7):727–+. pmid:25985139
  38. 38. Lichten M. Tetrad, random spore, and molecular analysis of meiotic segregation and recombination. In: Smith SJ, Burke JD, editors. Yeast Genetics: Methods and Protocols. New York, NY: Springer New York; 2014. p. 13–28.
  39. 39. Borde V, Robine N, Lin W, Bonfils S, Géli V, Nicolas A. Histone H3 lysine 4 trimethylation marks meiotic recombination initiation sites. The EMBO Journal. 2008;28(2):99–111. pmid:19078966
  40. 40. Pan J, Sasaki M, Kniewel R, Murakami H, Blitzblau Hannah G, Tischfield Sam E, et al. A hierarchical combination of factors shapes the genome-wide topography of yeast meiotic recombination rnitiation. Cell. 2011;144(5):719–31. pmid:21376234
  41. 41. Drouaud J, Khademian H, Giraut L, Zanni V, Bellalou S, Henderson IR, et al. Contrasted patterns of crossover and non-crossover at Arabidopsis thaliana meiotic recombination hotspots. PLoS Genetics. 2013;9(11).
  42. 42. Choi K, Zhao X, Kelly KA, Venn O, Higgins JD, Yelina NE, et al. Arabidopsis meiotic crossover hot spots overlap with H2A.Z nucleosomes at gene promoters. Nature Genetics. 2013;45(11):1327–36.—supplementary-information. pmid:24056716
  43. 43. Hellsten U, Wright KM, Jenkins J, Shu S, Yuan Y, Wessler SR, et al. Fine-scale variation in meiotic recombination in Mimulus inferred from population shotgun sequencing. Proceedings of the National Academy of Sciences USA. 2013;110(48):19478–82.
  44. 44. Muñoz-Fuentes V, Di Rienzo A, Vilà C. Prdm9, a major determinant of meiotic recombination hotspots, is not functional in dogs and their wild relatives, wolves and coyotes. PLoS ONE. 2011;6(11):e25498. pmid:22102853
  45. 45. Baudat F, Buard J, Grey C, Fledel-Alon A, Ober C, Przeworski M, et al. PRDM9 is a major determinant of meiotic recombination hotspots in humans and mice. Science. 2010;327(5967):836–40. pmid:20044539
  46. 46. Berg IL, Neumann R, Lam K- WG, Sarbajna S, Odenthal-Hesse L, May CA, et al. PRDM9 variation strongly influences recombination hot-spot activity and meiotic instability in humans. Nature Genetics. 2010;42(10):859–63.—supplementary-information. pmid:20818382
  47. 47. Myers S, Bowden R, Tumian A, Bontrop RE, Freeman C, MacFie TS, et al. Drive against hotspot motifs in primates implicates the PRDM9 gene in meiotic recombination. Science. 2010;327(5967):876–9. pmid:20044541
  48. 48. Parvanov ED, Petkov PM, Paigen K. Prdm9 controls activation of mammalian recombination hotspots. Science. 2010;327(5967):835–.
  49. 49. Hayashi K, Yoshida K, Matsui Y. A histone H3 methyltransferase controls epigenetic events required for meiotic prophase. Nature. 2005;438(7066):374–8. pmid:16292313
  50. 50. Walker M, Billings T, Baker CL, Powers N, Tian H, Saxl RL, et al. Affinity-seq detects genome-wide PRDM9 binding sites and reveals the impact of prior chromatin modifications on mammalian recombination hotspot usage. Epigenetics & Chromatin. 2015;8(1):1–13.
  51. 51. Smagulova F, Gregoretti IV, Brick K, Khil P, Camerini-Otero RD, Petukhova GV. Genome-wide analysis reveals novel molecular features of mouse recombination hotspots. Nature. 2011;472(7343):375–8.—supplementary-information. pmid:21460839
  52. 52. Myers S, Bottolo L, Freeman C, McVean G, Donnelly P. A fine-scale map of recombination rates and hotspots across the human genome. Science. 2005;310:321–4. pmid:16224025
  53. 53. Ponting CP. What are the genomic drivers of the rapid evolution of PRDM9? Trends in Genetics. 2011;27(5):165–71. pmid:21388701
  54. 54. Singhal S, Leffler EM, Sannareddy K, Turner I, Venn O, Hooper DM, et al. Stable recombination hotspots in birds. Science. 2015;350(6263):928–32. pmid:26586757
  55. 55. Kawakami T, Smeds L, Backström N, Husby A, Qvarnström A, Mugal CF, et al. A high-density linkage map enables a second-generation collared flycatcher genome assembly and reveals the patterns of avian recombination rate variation and chromosomal evolution. Molecular Ecology. 2014;23(16):4035–58. pmid:24863701
  56. 56. van Oers K, Santure AW, De Cauwer I, van Bers NE, Crooijmans RP, Sheldon BC, et al. Replicated high-density genetic maps of two great tit populations reveal fine-scale genomic departures from sex-equal recombination rates. Heredity. 2014;112(3):307–16. pmid:24149651
  57. 57. Backstrom N, Forstmeier W, Schielzeth H, Mellenius H, Nam K, Bolund E, et al. The recombination landscape of the zebra finch Taeniopygia guttata genome. Genome Research. 2010;20(4):485–95. pmid:20357052
  58. 58. White M. Animal Cytology and Evolution. London: Cambridge University Press; 1973.
  59. 59. Wang S, Zickler D, Kleckner N, Zhang L. Meiotic crossover patterns: Obligatory crossover, interference and homeostasis in a single process. Cell Cycle. 2015;14(3):305–14. pmid:25590558
  60. 60. Oliver PL, Goodstadt L, Bayes JJ, Birtle Z, Roach KC, Phadnis N, et al. Accelerated evolution of the Prdm9 speciation gene across diverse Metazoan taxa. PLoS Genetics. 2009;5(12):e1000753. pmid:19997497
  61. 61. Ellegren H, Smeds L, Burri R, Olason PI, Backstrom N, Kawakami T, et al. The genomic landscape of species divergence in Ficedula flycatchers. Nature. 2012;491(7426):756–60.—supplementary-information. pmid:23103876
  62. 62. La Smeds, Kawakami T, Burri R, Bolivar P, Husby A, Qvarnström A, et al. Genomic identification and characterization of the pseudoautosomal region in highly differentiated avian sex chromosomes. Nature Communications. 2014;5:5448. pmid:25378102
  63. 63. Petronczki M, Siomos MF, Nasmyth K. Un Ménage à Quatre: The Molecular Biology of Chromosome Segregation in Meiosis. Cell. 2003;112(4):423–40. pmid:12600308
  64. 64. Arbeithuber B, Betancourt AJ, Ebner T, Tiemann-Boege I. Crossovers are associated with mutation and biased gene conversion at recombination hotspots. Proceedings of the National Academy of Sciences USA. 2015;112(7):2109–14.
  65. 65. Stapley J, Birkhead TR, Burke T, Slate J. A linkage map of the zebra finch Taeniopygia guttata provides new insights into avian genome evolution. Genetics. 2008;179(1):651–67. pmid:18493078
  66. 66. Wang J, Fan HC, Behr B, Quake SR. Genome-wide single-cell analysis of recombination activity and de novo mutation rates in human sperm. Cell. 2012;150:402–12. pmid:22817899
  67. 67. McVean GAT, Myers SR, Hunt S, Deloukas P, Bentley DR, Donnelly P. The fine-scale structure of recombination rate variation in the human genome. Science. 2004;304(5670):581–4. pmid:15105499
  68. 68. Nishant KT, Rao MRS. Molecular features of meiotic recombination hot spots. BioEssays. 2006;28(1):45–56. pmid:16369948
  69. 69. Lichten M. Meiotic chromatin: The substrate for recombination initiation. In: Egel R, Lankenau D-H, editors. Recombination and meiosis: models, means, and evolution. Berlin, Germany: Springer-Verlag; 2008. p. 165–93.
  70. 70. Brunschwig H, Levi L, Ben-David E, Williams RW, Yakir B, Shifman S. Fine-scale maps of recombination rates and hotspots in the mouse genome. Genetics. 2012;191(3):757–64. pmid:22562932
  71. 71. Brick K, Smagulova F, Khil P, Camerini-Otero RD, Petukhova GV. Genetic recombination is directed away from functional genomic elements in mice. Nature. 2012;485(7400):642–5.—supplementary-information. pmid:22660327
  72. 72. Deaton AM, Bird A. CpG islands and the regulation of transcription. Genes & Development. 2011;25(10):1010–22.
  73. 73. Lichten M. Putting the breaks on meiosis. Science. 2015;350(6263):913–. pmid:26586748
  74. 74. Chan AH, Jenkins PA, Song YS. Genome-wide fine-scale recombination rate variation in Drosophila melanogaster. PLoS Genetics. 2012;8(12):e1003090. pmid:23284288
  75. 75. Smukowski Heil CS, Ellison C, Dubin M, Noor MAF. Recombining without hotspots: A comprehensive evolutionary portrait of recombination in two closely related species of Drosophila. Genome Biology and Evolution. 2015.
  76. 76. Kaur T, Rockman MV. Crossover hHeterogeneity in the absence of hotspots in Caenorhabditis elegans. Genetics. 2014;196(1):137–48. pmid:24172135
  77. 77. Wallberg A, Glemin S, Webster MT. Extreme recombination frequencies shape genome variation and evolution in the honeybee, Apis mellifera. PLoS Genetics. 2015;11(4).
  78. 78. Cole F, Baudat F, Grey C, Keeney S, de Massy B, Jasin M. Mouse tetrad analysis provides insights into recombination mechanisms and hotspot evolutionary dynamics. Nature Genetics. 2014;46(10):1072–80.—supplementary-information. pmid:25151354
  79. 79. Baker CL, Kajita S, Walker M, Saxl RL, Raghupathy N, Choi K, et al. PRDM9 drives evolutionary erosion of hotspots in Mus musculus through haplotype-specific initiation of meiotic recombination. PLoS Genetics. 2015;11(1):e1004916. pmid:25568937
  80. 80. Axelsson E, Webster MT, Ratnakumar A, The LC, Ponting CP, Lindblad-Toh K. Death of PRDM9 coincides with stabilization of the recombination landscape in the dog genome. Genome Research. 2012;22(1):51–63. pmid:22006216
  81. 81. Lam I, Keeney S. Nonparadoxical evolutionary stability of the recombination initiation landscape in yeast. Science. 2015;350(6263):932–7. pmid:26586758
  82. 82. Backström N, Karaiskou N, Leder EH, Gustafsson L, Primmer CR, Qvarnström A, et al. A gene-based genetic linkage map of the collared flycatcher (Ficedula albicollis) reveals extensive synteny and gene-order conservation during 100 million years of avian evolution. Genetics. 2008;179(3):1479–95. pmid:18562642
  83. 83. Lenormand T, Dutheil J. Recombination difference between sexes: a role for haploid selection. PLoS Biology. 2005;3(3):e63. pmid:15736976
  84. 84. Aslam M, Bastiaansen J, Crooijmans R, Vereijken A, Megens H-J, Groenen M. A SNP based linkage map of the turkey genome reveals multiple intrachromosomal rearrangements between the turkey and chicken genomes. BMC Genomics. 2010;11(1):647.
  85. 85. Hansson B, Åkesson M, Slate J, Pemberton JM. Linkage mapping reveals sex-dimorphic map distances in a passerine bird. Proceedings of the Royal Society of London B: Biological Sciences. 2005;272(1578):2289–98.
  86. 86. Kayang BB, Vignal A, Inoue-Murayama M, Miwa M, Monvoisin JL, Ito S, et al. A first-generation microsatellite linkage map of the Japanese quail. Animal Genetics. 2004;35(3):195–200. pmid:15147390
  87. 87. Haldane JBS. Sex ratio and unisexual sterility in hybrid animals. Journ of Gen. 1922;12(2):101–9.
  88. 88. Huxley JS. Sexual difference of linkage in Gammarus chevreuxi. Journ of Gen. 1928;20(2):145–56.
  89. 89. Trivers R. Sex differences in rates of recombination and sexual selection. In: Michod R, Levin B, editors. The evolution of sex. Sunderland, MA: Sinauer; 1988. p. 270–86.
  90. 90. Stevison LS. Male-mediated effects on female meiotic recombination. Evolution. 2012;66(3):905–11. pmid:22380449
  91. 91. Mank JE. The evolution of heterochiasmy: the role of sexual selection and sperm competition in determining sex-specific recombination rates in eutherian mammals. Genetics Research. 2009;91(05):355–63.
  92. 92. Broman KW, Murray JC, Sheffield VC, White RL, Weber JL. Comprehensive human genetic maps: individual and sex-specific variation in recombination. American Journal of Human Genetics. 1998;63(3):861–9. pmid:9718341
  93. 93. Paigen K, Petkov P. Mammalian recombination hot spots: properties, control and evolution. Nature Reviews Genetics. 2010;11(3):221–33. pmid:20168297
  94. 94. Hillers KJ. Crossover interference. Current Biology. 2004;14:R1036–7. pmid:15620632
  95. 95. Segura J, Ferretti L, Ramos-Onsins S, Capilla L, Farré M, Reis F, et al. Evolution of recombination in eutherian mammals: insights into mechanisms that affect recombination rates and crossover interference. Proceedings of the Royal Society of London B: Biological Sciences. 2013;280(1771).
  96. 96. Petkov PM, Broman KW, Szatkiewicz JP, Paigen K. Crossover interference underlies sex differences in recombination rates. Trends in Genetics. 2007;23(11):539–42. pmid:17964681
  97. 97. Cutter AD, Payseur BA. Genomic signatures of selection at linked sites: unifying the disparity among species. Nature Reviews Genetics. 2013;14(4):262–74. pmid:23478346
  98. 98. Odenthal-Hesse L, Berg IL, Veselis A, Jeffreys AJ, May CA. Transmission distortion affecting human noncrossover but not crossover recombination: A hidden source of meiotic drive. PLoS Genetics. 2014;10(2):e1004106. pmid:24516398
  99. 99. Williams AL, Genovese G, Dyer T, Altemose N, Truax K, Jun G, et al. Non-crossover gene conversions show strong GC bias and unexpected clustering in humans. eLife. 2015;4:e04637.
  100. 100. Nadachowska-Brzyska K, Burri R, Olason PI, Kawakami T, Smeds L, Ellegren H. Demographic divergence history of pied Flycatcher and collared flycatcher inferred from whole-genome re-sequencing data. PLoS Genetics. 2013;9(11):e1003942. pmid:24244198
  101. 101. Mugal CF, Arndt PF, Ellegren H. Twisted signatures of GC-biased gene conversion embedded in an evolutionary stable karyotype. Molecular Biology and Evolution. 2013;30(7):1700–12. pmid:23564940
  102. 102. Bolívar P, Mugal CF, Nater A, Ellegren H. Recombination rate variation modulates gene sequence evolution mainly via GC-biased gene conversion, not Hill-Robertson interference, in an avian system. Molecular Biology and Evolution. 2015.
  103. 103. Li H, Durbin R. Fast and accurate short read alignment with Burrows-Wheeler Transform. Bioinformatics. 2009;25:1754–60. pmid:19451168
  104. 104. DePristo MA, Banks E, Poplin R, Garimella KV, Maguire JR, Hartl C, et al. A framework for variation discovery and genotyping using next-generation DNA sequencing data. Nature Genetics. 2011;43(5):491–8.—supplementary-information. pmid:21478889
  105. 105. McKenna A, Hanna M, Banks E, Sivachenko A, Cibulskis K, Kernytsky A, et al. The Genome Analysis Toolkit: A MapReduce framework for analyzing next-generation DNA sequencing data. Genome Research. 2010;20(9):1297–303. pmid:20644199
  106. 106. Van der Auwera GA, Carneiro MO, Hartl C, Poplin R, del Angel G, Levy-Moonshine A, et al. From FastQ data to high-confidence variant calls: the Genome Analysis Toolkit best practices pipeline. Current Protocols in Bioinformatics: John Wiley & Sons, Inc.; 2002.
  107. 107. Kawakami T, Backström N, Burri R, Husby A, Olason P, Rice AM, et al. Estimation of linkage disequilibrium and interspecific gene flow in Ficedula flycatchers by a newly developed 50k single-nucleotide polymorphism array. Molecular Ecology Resources. 2014;14(6):1248–60. pmid:24784959
  108. 108. Smeds L, Warmuth V, Bolivar P, Uebbing S, Burri R, Suh A, et al. Evolutionary analysis of the female-specific avian W chromosome. Nature Communications. 2015;6:7330. pmid:26040272
  109. 109. Benson G. Tandem repeats finder: a program to analyze DNA sequences. Nucleic Acids Research. 1999;27(2):573–80. pmid:9862982
  110. 110. Thorvaldsdóttir H, Robinson JT, Mesirov JP. Integrative Genomics Viewer (IGV): high-performance genomics data visualization and exploration. Briefings in Bioinformatics. 2013;14(2):178–92. pmid:22517427
  111. 111. Hackenberg M, Previti C, Luque-Escamilla P, Carpena P, Martinez-Aroza J, Oliver J. CpGcluster: a distance-based algorithm for CpG-island detection. BMC Bioinformatics. 2006;7(1):446.