Bayesian population genomic inference of crossing over and gene conversion

Meiotic recombination is a fundamental cellular mechanism in sexually reproducing organisms and its different forms, crossing over and gene conversion both play an important role in shaping genetic variation in populations. Here, we describe a coalescent-based full-likelihood Markov chain Monte Carlo (MCMC) method for jointly estimating the crossing-over, gene-conversion, and mean tract length parameters from population genomic data under a Bayesian framework. Although computationally more expensive than methods that use approximate likelihoods, the relative efficiency of our method is expected to be optimal in theory. Furthermore, it is also possible to obtain a posterior sample of genealogies for the data using this method. We first check the performance of the new method on simulated data and verify its correctness. We also extend the method for inference under models with variable gene-conversion and crossing-over rates and demonstrate its ability to identify recombination hotspots. Then, we apply the method to two empirical data sets that were sequenced in the telomeric regions of the X chromosome of Drosophila melanogaster. Our results indicate that gene conversion occurs more frequently than crossing over in the su-w and su-s gene sequences while the local rates of crossing over as inferred by our program are not low. The mean tract lengths for gene-conversion events are estimated to be ～70 bp and 430 bp, respectively, for these data sets. Finally, we discuss ideas and optimizations for reducing the execution time of our algorithm.     

Estimating recombination rates from single-nucleotide polymorphisms using summary statistics

We describe a novel method for jointly estimating crossing-over and gene-conversion rates from population genetic data using summary statistics. The performance of our method was tested on simulated data sets and compared with the composite-likelihood method of R. R. Hudson. For several realistic parameter values, the new method performed similarly to the composite-likelihood approach for estimating crossing-over rates and better when estimating gene-conversion rates. We used our method to analyze a human data set recently genotyped by Perlegen Sciences.     

Causes and consequences of crossing-over evidenced via a high-resolution recombinational landscape of the honey bee

BackgroundSocial hymenoptera, the honey bee (Apis mellifera) in particular, have ultra-high crossover rates and a large degree of intra-genomic variation in crossover rates. Aligned with haploid genomics of males, this makes them a potential model for examining the causes and consequences of crossing over. To address why social insects have such high crossing-over rates and the consequences of this, we constructed a high-resolution recombination atlas by sequencing 55 individuals from three colonies with an average marker density of 314 bp/marker.ResultsWe find crossing over to be especially high in proximity to genes upregulated in worker brains, but see no evidence for a coupling with immune-related functioning. We detect only a low rate of non-crossover gene conversion, contrary to current evidence. This is in striking contrast to the ultrahigh crossing-over rate, almost double that previously estimated from lower resolution data. We robustly recover the predicted intragenomic correlations between crossing over and both population level diversity and GC content, which could be best explained as indirect and direct consequences of crossing over, respectively.ConclusionsOur data are consistent with the view that diversification of worker behavior, but not immune function, is a driver of the high crossing-over rate in bees. While we see both high diversity and high GC content associated with high crossing-over rates, our estimate of the low non-crossover rate demonstrates that high non-crossover rates are not a necessary consequence of high recombination rates.

Electronic supplementary material

The online version of this article (doi:10.1186/s13059-014-0566-0) contains supplementary material, which is available to authorized users.     

Gene conversion in the Escherichia coli RecF pathway: a successive half crossing-over model.

Summary Gene conversion - apparently non-reciprocal transfer of sequence information between homologous DNA sequences - has been reported in various organisms. Frequent association of gene conversion with reciprocal exchange (crossing-over) of the flanking sequences in meiosis has formed the basis of the current view that gene conversion reflects events at the site of interaction during homologous recombination. In order to analyze mechanisms of gene conversion and homologous recombination in an Escherichia coli strain with an active RecF pathway (recBC sbcBC), we first established in cells of this strain a plasmid carrying two mutant neo genes, each deleted for a different gene segment, in inverted orientation. We then selected kanamycin-resistant plasmids that had reconstituted an intact neo ⁺ gene by homologous recombination. We found that all the neo ⁺ plasmids from these clones belonged to the gene-conversion type in the sense that they carried one neo ⁺ gene and retained one of the mutant neo genes. This apparent gene conversion was, however, only very rarely accompanied by apparent crossing-over of the flanking sequences. This is in contrast to the case in a rec ⁺ strain. or in a strain with an active RecE pathway (recBC sbcA). Our further analyses, especially comparisons with apparent gene conversion in the rec ⁺ strain, led us to propose a mechanism for this biased gene conversion. This successive half crossing-over model proposes that the elementary recombinational process is half crossing;-over in the sense that it generates only one recombinant DNA duplex molecule, and leaves one or two free end(s), out of two parental DNA duplexes. The resulting free end is, the model assumes, recombinogenic and frequently engages in a second round of half crossing-over with the recombinant duplex. The products resulting from such interaction involving two molecules of the plasmid would be classified as belonging to the gene-conversion type without crossing-over. We constructed a dimeric molecule that mimics the intermediate form hypothesized in this model and introduced it into cells. Biased gene conversion products were obtained in this reconstruction experiment. The half crossing-over mechanism can also explain formation of huge linear multimers of bacterial plasmids, the nature of transcribable recombination products in bacterial conjugation, chromosomal gene conversion not accompanied by flanking exchange (like that in yeast mating-type switching), and antigenic variation in microorganisms.     

Linkage disequilibrium patterns across a recombination gradient in African Drosophila melanogaster

Previous multilocus surveys of nucleotide polymorphism have documented a genome-wide excess of intralocus linkage disequilibrium (LD) in Drosophila melanogaster and D. simulans relative to expectations based on estimated mutation and recombination rates and observed levels of diversity. These studies examined patterns of variation from predominantly non-African populations that are thought to have recently expanded their ranges from central Africa. Here, we analyze polymorphism data from a Zimbabwean population of D. melanogaster, which is likely to be closer to the standard population model assumptions of a large population with constant size. Unlike previous studies, we find that levels of LD are roughly compatible with expectations based on estimated rates of crossing over. Further, a detailed examination of genes in different recombination environments suggests that markers near the telomere of the X chromosome show considerably less linkage disequilibrium than predicted by rates of crossing over, suggesting appreciable levels of exchange due to gene conversion. Assuming that these populations are near mutation-drift equilibrium, our results are most consistent with a model that posits heterogeneity in levels of exchange due to gene conversion across the X chromosome, with gene conversion being a minor determinant of LD levels in regions of high crossing over. Alternatively, if levels of exchange due to gene conversion are not negligible in regions of high crossing over, our results suggest a marked departure from mutation-drift equilibrium (i.e., toward an excess of LD) in this Zimbabwean population. Our results also have implications for the dynamics of weakly selected mutations in regions of reduced crossing over.     

Recombinogenic effects of suppressors of position-effect variegation in Drosophila

Compact chromatin structure, induction of gene silencing in position-effect variegation (PEV), and crossing-over suppression are typical features of heterochromatin. To identify genes affecting crossing-over suppression by heterochromatin we tested PEV suppressor mutations for their effects on crossing over in pericentromeric regions of Drosophila autosomes. From the 46 mutations (28 loci) studied, 16 Su(var) mutations of the nine genes Su(var)2-1, Su(var)2-2, Su(var)2-5, Su(var)2-10, Su(var)2-14, Su(var)2-15, Su(var)3-3, Su(var)3-7, and Su(var)3-9 significantly increase in heterozygotes or by additive effects in double and triple heterozygotes crossing over in the ri-p(p) region of chromosome 3. Su(var)2-2(01) and Su(var)2-14(01) display the strongest recombinogenic effects and were also shown to enhance recombination within the light-rolled heterochromatic region of chromosome 2. The dominant recombinogenic effects of Su(var) mutations are most pronounced in proximal euchromatin and are accompanied with significant reduction of meiotic nondisjunction. Our data suggest that crossing-over suppression by heterochromatin is controlled at chromatin structure as well as illustrate the possible effects of heterochromatin on total crossing-over frequencies in the genome.     

Relative influences of crossing over and gene conversion on the pattern of linkage disequilibrium in Arabidopsis thaliana

In this article we infer the rates of gene conversion and crossing over in Arabidopsis thaliana from population genetic data. Our data set is a genomewide survey consisting of 1347 fragments of length 600 bp sequenced in 96 accessions. It has several orders of magnitude more markers than any previous nonhuman study. This allows for more accurate inference as well as a detailed comparison between theoretical expectations and observations. Our methodology is specifically set to account for deviations such as recurrent mutations or a skewed frequency spectrum. We found that even if some components of the model clearly do not fit, the pattern of LD conforms to theoretical expectations quite well. The ratio of gene conversion to crossing over is estimated to be around one. We also find evidence for fine-scale variations of the crossing-over rate.     

A 6-bp deletion 5'' to the G gamma globin gene in beta S chromosomes bearing the Bantu haplotype.

Sequencing of the upstream region of a human G gamma gene linked to the Bantu haplotype revealed a 6-bp deletion between site -400 and -395. Further analysis revealed that this mutation is present in 37% of the sickle cell anemia patients bearing the Bantu haplotype and is absent in the other haplotypes linked to the beta S gene, as well as in most chromosomes bearing the beta A-globin gene. The most parsimonious interpretation of the data is that the deletion is a very recent event which occurred in the subset of Bantu chromosomes already bearing a gene conversion of the A gamma gene by the G gamma gene. Its presence in black beta S chromosomes is most probably the consequence of a crossing-over between a Bantu beta S chromosome (with deletion and gene conversion) and a beta A chromosome.     

Estimating the contribution of mutation, recombination and gene conversion in the generation of haplotypic diversity

Recombination occurs through both homologous crossing over and homologous gene conversion during meiosis. The contribution of recombination relative to mutation is expected to be dramatically reduced in inbreeding organisms. We report coalescent-based estimates of the recombination parameter (rho) relative to estimates of the mutation parameter (theta) for 18 genes from the highly self-fertilizing grass, wild barley, Hordeum vulgare ssp. spontaneum. Estimates of rho/theta are much greater than expected, with a mean rho/theta approximately 1.5, similar to estimates from outcrossing species. We also estimate rho with and without the contribution of gene conversion. Genotyping errors can mimic the effect of gene conversion, upwardly biasing estimates of the role of conversion. Thus we report a novel method for identifying genotyping errors in nucleotide sequence data sets. We show that there is evidence for gene conversion in many large nucleotide sequence data sets including our data that have been purged of all detectable sequencing errors and in data sets from Drosophila melanogaster, D. simulans, and Zea mays. In total, 13 of 27 loci show evidence of gene conversion. For these loci, gene conversion is estimated to contribute an average of twice as much as crossing over to total recombination.     

Double crossing over: elementary events and the sequence of their occurrence

Analysis of the crossing over increment in the structurally normal chromosome of Drosophila caused by a rearrangement in nonhomologous chromosome (interchromosomal effect on crossing over, IEC) was carried out based on the author's personal and literature data. The IEC in the left arm of chromosome 2 caused by inversions in chromosomes X and 3, as well as the IEC in X chromosome caused by inversions in chromosomes 2 and 3, were examined. The IEC-induced increment of crossing over results from the increase of the number of double exchanges under the constant or reduced number of single exchanges. Tetrad analysis showed that the given alternation of the crossing over processes could occur only in the case of conversion of the tetrads with single exchanges into the tetrads with double exchanges. In other words, the events leading to the formation of double exchanges occur consecutively. The borders of the IEC-induced double exchanges can be seen all over the chromosome body. However, the IEC-induced increase of chromosome recombination length occurs only in the proximal region (in rare cases, in proximal and distal regions) of the chromosome arm. This means that a double exchange is formed when the first event with predominant location in the middle of the arm is supplemented with the second event predominantly localized at the arm T end, most frequently in the proximal region. The pattern of the IEC-induced double exchange formation can be satisfactorily described in terms of the contact model of the crossing over. According to the model, an elementary crossing-over event is the local contact between the homologues. Neither single exchange nor a double-stranded DNA break can serve as an elementary event in the process of any multiple exchange formation.     

Estimating recombination rates using three-site likelihoods

We introduce a new method for jointly estimating crossing-over and gene conversion rates using sequence polymorphism data. The method calculates probabilities for subsets of the data consisting of three segregating sites and then forms a composite likelihood by multiplying together the probabilities of many subsets. Simulations show that this new method performs better than previously proposed methods for estimating gene conversion rates, but that all methods require large amounts of data to provide reliable estimates. While existing methods can easily estimate an "average" gene conversion rate over many loci, they cannot reliably estimate gene conversion rates for a single region of the genome.     

Mitotic intragenic recombination as a consequence of heteroduplex formation in Aspergillus nidulans

Summary A diploid strain of Aspergillus nidulans with two heteroallelic mutations in the pabaA cistron (right arm of the first chromosome) has been studied. Part of the paba-independent colonies which have been examined was heterogeneous, i.e. they showed conidia of different colour and genotype. The genetic analysis of the various type of these heterogeneous colonies leads to the conclusion that, in Aspergillus nidulans, mitotic intragenic recombination is, in most cases, consequence of a single-strand break and exchange followed by the formation of a very long hybrid-DNA region (in our case a maximum of 22 meiotic units); the selected characteristics arise mainly by gene-conversion.Furthermore, data show a high negative interference between the selected crossing-over and a second crossing-over on the left arm and probably also on different chromosomes. The latter exchange occurs, as the former, between subchromatidic units.     

Crossing over in chicken oogenesis: cytological and chiasma-based genetic maps of the chicken lampbrush chromosome 1

Chiasmata in diplotene bivalents are located at the points of physical exchange (crossing-over) between homologous chromosomes. We have studied chiasma distribution within chicken lampbrush chromosome 1 to estimate the crossing-over frequency between chromosome landmarks. The position of the centromere and chromosome region 1q3.3-1q3.6 on lampbrush chromosome 1 were determined by comparative physical mapping of the TTAGGG repeats in the chicken mitotic and lampbrush chromosomes. The comparison of the chiasma (=crossing over)-based genetic distances on chicken chromosome 1 with the genetic linkage map obtained in genetic experiments showed that current genetic distances estimated by the high-resolution genetic mapping of the East Lansing, Compton, and Wageningen chicken reference populations are 1.2-1.9 times longer than those based on chiasma counts. Conceivable reasons for this discrepancy are discussed.     

Molecular analysis of a hybrid gene encoding human glycophorin variant Miltenberger V-like molecule

The genomic structure of a human glycophorin variant, Miltenberger class V-like molecule (MiV*), was examined. Southern blot analysis of total genomic DNA revealed that the 5' half of the MiV* gene derived from glycophorin A (GPA) gene whereas the 3' half derived from glycophorin B (GPB) gene. This structure is reciprocal to another glycophorin variant, Sta, which has a GPB-GPA hybrid structure. The genomic sequences around the crossing-over point were amplified by polymerase chain reaction, and the sequences were determined. Comparison of the nucleotide sequences of the GPA, GPB, and MiV* genes indicates that the crossing-over point is located in the region around the 3' end of intron 3 of the GPA gene. This place is different from the crossing-over point for Sta, which was found to be highly homologous to that for haptoglobin-related genes. However, the nucleotide sequences within the presumptive crossing-over point for the MiV* gene were found to be homologous in a reverse orientation to the crossing-over point proposed for haptoglobin-related genes. These results suggest strongly that homologous recombination through unequal crossing over can be facilitated by specific genomic elements such as those in common for formation of MiV*, Sta, and haptoglobin-related genes. The present study also localized the gene of the third glycophorin, GPE, at chromosome 4, q31.1 band, the same locus as for the GPA and GPB genes. The results indicate that GPE was not involved in generating MiV* or Sta hybrid gene despite the fact that it is localized adjacent to the GPA and GPB genes.     

Peakwide mapping on chromosome 3q13 identifies the kalirin gene as a novel candidate gene for coronary artery disease

A susceptibility locus for coronary artery disease (CAD) has been mapped to chromosome 3q13-21 in a linkage study of early-onset CAD. We completed an association-mapping study across the 1-LOD-unit-down supporting interval, using two independent white case-control data sets (CATHGEN, initial and validation) to evaluate association under the peak. Single-nucleotide polymorphisms (SNPs) evenly spaced at 100-kb intervals were screened in the initial data set (N=468). Promising SNPs (P<.1) were then examined in the validation data set (N=514). Significant findings (P<.05) in the combined initial and validation data sets were further evaluated in multiple independent data sets, including a family-based data set (N=2,954), an African American case-control data set (N=190), and an additional white control data set (N=255). The association between genotype and aortic atherosclerosis was examined in 145 human aortas. The peakwide survey found evidence of association in SNPs from multiple genes. The strongest associations were found in three SNPs from the kalirin (KALRN) gene, especially in patients with early-onset CAD (P=.00001-00028 in the combined CATHGEN data sets). In-depth investigation of the gene found that an intronic SNP, rs9289231, was associated with early-onset CAD in all white data sets examined (P<.05). In the joint analysis of all white early-onset CAD cases (N=332) and controls (N=546), rs9289231 was highly significant (P=.00008), with an odds-ratio estimate of 2.1. Furthermore, the risk allele of this SNP was associated with atherosclerosis burden (P=.03) in 145 human aortas. KALRN is a protein with many functions, including the inhibition of inducible nitric oxide synthase and guanine-exchange-factor activity. KALRN and two other associated genes identified in this study (CDGAP and MYLK) belong to the Rho GTPase-signaling pathway. Our data suggest the importance of the KALRN gene and the Rho GTPase-signaling pathway in the pathogenesis of CAD.     

Identification and characterization of coding single-nucleotide polymorphisms within a human olfactory receptor gene cluster

Single-nucleotide polymorphisms (SNPs) were studied in 15 olfactory receptor (OR) coding regions, one control region and two noncoding sequences all residing within a 412 kb OR gene cluster on human chromosome 17p13.3, as well as in other G-protein coupled receptors (GPCRs). A total of 26 SNPs were identified in ORs, 21 of which are coding SNPs (cSNPs). The mean nucleotide diversity of OR coding regions was 0.078% (ranging from 0 to 0.16%), which is about twice higher than that of other GPCRs, and similar to the nucleotide diversity levels of noncoding regions along the human genome. The high polymorphism level in the OR coding regions might be due to a weak positive selection pressure acting on the OR genes. In two cases, OR genes have been found to share the same cSNP. This could be explained by recent gene conversion events, which might be a part of a concerted evolution mechanism acting on the OR superfamily. Using the genotype data of 85 unrelated individuals in 15 SNPs, we found linkage disequilibrium (LD) between pairs of SNPs located on the centromeric part of the cluster. On the other hand, no LD was found between SNPs located on the telomeric part of the cluster, suggesting the presence of several hot-spots for recombination within this cluster. Thus, different regions of this gene cluster may have been subject to different recombination rates.     

Recombination and spontaneous mutation at the major cluster of resistance genes in lettuce (Lactuca sativa)

Two sets of overlapping experiments were conducted to examine recombination and spontaneous mutation events within clusters of resistance genes in lettuce. Multiple generations were screened for recombinants using PCR-based markers flanking Dm3. The Dm3 region is not highly recombinagenic, exhibiting a recombination frequency 18-fold lower than the genome average. Recombinants were identified only rarely within the cluster of Dm3 homologs and no crossovers within genes were detected. Three populations were screened for spontaneous mutations in downy mildew resistance. Sixteen Dm mutants were identified corresponding to spontaneous mutation rates of 10(-3) to 10(-4) per generation for Dm1, Dm3, and Dm7. All mutants carried single locus, recessive mutations at the corresponding Dm locus. Eleven of the 12 Dm3 mutations were associated with large chromosome deletions. When recombination could be analyzed, deletion events were associated with exchange of flanking markers, consistent with unequal crossing over; however, although the number of Dm3 paralogs was changed, no novel chimeric genes were detected. One mutant was the result of a gene conversion event between Dm3 and a closely related homolog, generating a novel chimeric gene. In two families, spontaneous deletions were correlated with elevated levels of recombination. Therefore, the short-term evolution of the major cluster of resistance genes in lettuce involves several genetic mechanisms including unequal crossing over and gene conversion.     

mus309 mutation, defective in DNA double-strand break repair, affects intergenic but not intragenic meiotic recombination in Drosophila melanogaster

The effect was investigated of the hypomorphic DNA double-strand break repair, notably synthesis-dependent strand annealing, deficient mutation mus309 on the third chromosome of Drosophila melanogaster on intergenic and intragenic meiotic recombination in the X chromosome. The results showed that the mutation significantly increases the frequency of intergenic crossing over in two of three gene intervals of the X chromosome studied. Interestingly the increase was most prevalent in the tip of the X chromosome where crossovers normally are least frequent per physical map unit length. In particular crossing over interference was also affected, indicating that the effect of the mus309 mutation involves preconditions of crossing over but not the event of crossing over itself. On the other hand, the results also show that most probably the mutation does not have any effect on intragenic recombination, i.e. gene conversion. These results are fully consistent with the present molecular models of meiotic crossing over initiated by double-strand breaks of DNA followed by formation of a single-end-invasion intermediate, or D-loop, which is subsequently processed to generate either crossover or non-crossover products involving formation of a double Holliday junction. In particular the results suggest that the mus309 gene is involved in resolution of the D-loop, thereby affecting the choice between double-strand-break repair (DSBR) and synthesis-dependent strand annealing (SDSA) pathways of meiotic recombination.