This study constitutes the first mapping effort of the Siberian jay genome, and is among the first ones to present a preliminary linkage map for any entirely natural vertebrate species from the wild (reviewed in ). The linkage map was composed using 107 polymorphic microsatellite loci typed on ca. 350 animals, making it one of the most detailed linkage maps available for natural animal populations . In fact, it is one of only a few genetic linkage maps of wild bird species to date. Apart from the revealing evidence for sex differences in recombination rates, the constructed maps represent an excellent resource from which the markers may be selected for future mapping projects in this and related species, as well as for comparative genomic studies of genome organisation. In what follows, we will discuss the salient features of the constructed linkage maps in comparison to similar maps and results from earlier studies. In particular, we will pay attention to sex-specific differences in recombination rates, map coverage and some other issues deserving future attention.
Genotyping in the mapping population
The constructed map contains 107 microsatellites, of which 101 are autosomal, three Z-chromosome-specific and three pseudoautosomal loci (see below). The ideal set of molecular marker data for linkage mapping has no missing values, no genotyping errors and the markers segregate in the expected ratio for the specific type of population . In practice, however, mapping data is compromised in all of these respects. However, as simulated and concluded in previous research , the effect of missing genotypes depends greatly on the sample size: the smaller the sample, the more severe the effects are likely to be. In comparison to published simulations, in which the results were found to be quite robust with 150 individuals and 10% of missing values, our data had much less missing values (4.5%) and a larger sample size (349 individuals). Hence, the potential biases in our map due to missing values are unlikely to be severe.
Map construction and genome coverage
The resulting linkage map of 107 microsatellites spans a total sex-average length of 886.6 cM, with 10 LGs. The number of markers within the data that showed significant linkage (LOD > 3.0) with at least one other marker was very high (107/117, 91.5%), a phenomenon observed also in a linkage mapping study of the great reed warbler . Of the 10 unlinked markers (LOD < 3.0), four demonstrate sufficiently informative meiosis, suggesting that these four unlinked markers were most likely located on unique chromosomes or chromosome arms. The haploid number of ca. 40 chromosomes is typical for passerine birds (e.g. 7 macro-, 32 micro-, and a pair of sex-chromosomes in the zebra finch genome ) and the chicken genome is composed of 39 haploid chromosomes (8 macro and 30 micro, and a pair of sex chromosomes ). Thus, the presence of small groups and unlinked markers indicates that appreciable gaps of at least 29 additional linkage groups should be filled to consolidate the Siberian jay map. The discrepancy between the number of LGs and the haploid number of chromosomes has been commonly reported when constructing linkage maps in avian species (see ). These results could be explained in terms of the non-random distribution of microsatellites in the avian genome, where microchromosomes – typically scarce of microsatellites  – constitute a large proportion (ca. 80%) of the total number of chromosomes.
Considering the avian order Passeriformes, the current sex-pooled map of the Siberian jay at 887 cM is larger than recent maps in the great reed warbler (155 – 237 cM ; 707 – 858 cM ), but smaller than those in the collared flycatcher (1787 cM ) and the zebra finch (1068 cM ). The maximum genome coverage of the markers described in this study was estimated to be 1187 cM (887 + 200 + 100 cM) (see[56, 57]), covering about a third of the Siberian jay genome of ~3800 cM, if estimated by assuming a similar genome size as in chicken . It is clear that some of the microchromosomes are poorly represented, or not represented at all, in the current map. This poorer coverage of the microchromosomes, as well as of the Z-chromosome, is in good agreement with the observations in chicken , Japanese quail , duck , and great reed warbler [4, 6]. However, it was argued in  that for the whole-genome map of the zebra finch using 876 SNPs the shorter length of linkage map relative to that of the chicken is ascribed to its unusually lower recombination rates. Similarly, lower recombination rates and smaller length of linkage map were also revealed in the collared flycatchers based on 147 gene markers and 64 microsatellites . Since it has been suggested that passerines generally have lower recombination rates than the chicken , the small map sizes here may also be partly due to reduced recombination rates in the Siberian jays. Nevertheless, the first-generation linkage map will undoubtedly evolve as more markers are added, with some additional linkage groups forming and some pairs of now described linkage groups coalescing into a single group. As linkage maps continue to develop, future work on the genetic map will increase the genome coverage by adding more novel genetic markers, for example, 1000's of SNP loci and additional markers will improve the resolving power of the map. A more saturated map should give more information about the genome size, exact karyotype number and chromosomal rearrangement of Siberian jay.
The constructed framework map was comprised only of loci with unambiguous positions relative to each other. As shown in Table 2, only a slight expansion of the framework map size was detected when the non-framework markers were included in the analysis. Non-significant (P > 0.05) difference in the interval distances between framework markers was found in the sex-specific maps when non-framework markers were included or not, which can be due to the fact that most (19/26) of the non-framework loci still can be placed with significant (LOD = 2; odds of 100:1) support in either of two alternative intervals at LOD = 3 (odds of 1000:1). However, on the sex-average maps, a significant (P < 0.05) difference in distances between framework markers was revealed when including or excluding the non-framework markers from the analysis. This is probably due to the overall effect of the recombination heterogeneity between the sexes on estimates of sex-average map distances between framework and non-framework markers . Thus, our results indicate that the framework maps can serve as the backbone of the best-position maps with a high level of confidence. Moreover, the framework map should also therefore provide a robust basis with which to apply to other pedigrees and for comparing genomic rearrangements with closely related species.
By assigning six genetic markers to the Z-chromosome on the basis of Mendelian inheritance and additional two-point analysis tests, we found that the recombination rates in the Z-chromosome linkage group differed substantially among the sexes. In particular the ratio of recombination fraction in females (θ
F) and males (θ
M, in the LGZ (θ
M < 1) contrasted markedly with those in seven of the nine autosomal linkage groups (θ
M > 1). These results conform to the Z-chromosome linkage results from the chicken  and the great reed warbler [4, 6], suggesting significant sex-specific heterogeneity in recombination rate on the avian Z-chromosome. Between markers SJ046 and SJ069, a small amount of recombination (recombination fraction θ = 0.02–0.05) was found to occur between Z and W chromosomes in the mapping pedigree. Z-W recombination has only once been previously reported in previous mapping efforts of birds based on SNPs (see ), but equivalent X-Y recombination has been observed in linkage mapping studies of sex chromosomes in many other species such as in salmonid fishes , rainbow trout , cattle , sheep [51, 62], oyster , and humans . However, in contrast to increased recombination rates in heterogametic sex (male, XY) as observed in the pseudoautosomal regions of above mentioned studies, an overall 8-fold lower recombination rates between SJ048 and SJ069 were found in female (ZW) as compared to male (ZZ) Siberian jays. In addition, it seems as the pseudoautosomal region is extensive in Siberian jay (15.8 cM/23.8 cM, 66.4%). While in other species, 13.3% (20.1 cM/151 cM) were reported for the PAR in the bovine sex-average consensus map , 6.3% (3.5 cM/55.8 cM) in the male-specific and 45.1% (54.2 cM/121 cM) in the female-specific map of the sex chromosomes in sheep , and a very small proportion of PAR in the chicken sex chromosomal maps . It is unclear why these are so, but potential explanations include the evolutionary strata , patterns of genome variability , distribution of sex-biased loci [67, 68], chiasma interference  and selection  on the avian Z-chromosome. Since the difference was found in the recombination rates and the change of genome structure during the avian evolution between Siberian jay and other wild bird species, we would like to speculate that the pattern of genome variability and distribution of sex-biased loci may have contributed more to this observation. Further investigations using more pseudoautosomal markers are needed to verify this result and understand its origin and significance.
When testing for sex-specific differences in recombination rates in different linkage groups and over the total autosomal map, we found evidence for statistically significant distortions towards reduced male recombination fractions only in two autosomal LGs, rather than in each of the nine autosomal LGs. This indicates that the sex-related differences in recombination rates are confined to certain, specific parts of the genome. This result is not unexpected given that in many species there is a large variation in the recombination rates within and among chromosomes (see [71, 72]). Similar situations of linkage group sex-specific distortion have also been observed in other species for example in marsupials (e.g. [73, 74]) and in various aquatic organisms (the pacific oyster ; the tiger pufferfish ; and the turbot ).
Sex-specific differences in recombination rates have been found in a diverse range of organisms from molluscs (e.g. ) and fish (e.g. ) to mammals (e.g. pig ; cattle ; and human ). However, in birds various patterns have been reported: little evidence of heterochiasmy in chicken (Gallus domesticus, ) and zebra finch (Taeniopygia guttata ); higher rates of recombination in males than in females in linkage maps of blue tit (Parus caeruleus ) and collared flycatcher (Ficedula albicollis ); and higher rates of recombination in females than in males in maps of great tit (Parus major ), great reed warbler (Acrocephalus arundinaceus [4, 6]) and house sparrow (Passer domesticus ). The sex-bias (females > males) observed in Siberian jay conforms to the last pattern opposing the Haldane's prediction  according to which the heterogametic sex should show lower recombination rates than the homogametic sex. So far the comparative data suggest a divergence between the genetic maps of the chicken (higher recombination rate and little difference in recombination rate between sexes) and the passerines (lower recombination rate and significant difference in recombination rate between sexes). Interestingly, on the one hand, we note the pronouncedly different patterns of recombination observed in collared flycatcher , zebra finch  and Siberian jay here, all of which are passerine birds. It has been proposed that heterochiasmy is the result of sexual selection, with the sex experiencing the greater variance in reproductive success exhibiting the lower recombination rate . However, this explanation may not be relevant here because both collared flycatchers  and zebra finches  are polygamous, whereas the Siberian jay is a monogamous species [34, 37], so that the reproductive success should be very similar between the sexes. On the other hand, we found that the female:male recombination rate (1.21) in the best-order map was much smaller than that observed in the great reed warbler where it varied from 2.15 (microsatellites) to 1.86 (AFLP markers ). This difference is not necessarily biologically meaningful as it could be attributable to less informative and smaller number of loci (albeit larger number of individuals) in the earlier studies. Furthermore, since marker density in both Siberian jay and great reed warbler maps is relatively low and the sex-specific recombination rates heterogenous among and within linkage groups, the estimated female:male recombination rates among the maps may vary if different genomic regions are mapped
The observed sex-specific recombination rates are potentially influenced by numerous factors, and at the moment, there is no consensus in respect to the relative importance of mechanisms accounting for the recombination differences (see [63, 73]). For example, numbers of hypotheses including sexual selection ), haploid selection , sex differences in the internal or external environment  and sex differences in gene expression  have been evoked to explain the heterochiasmy pattern in birds. In this long-term isolated population studied here, we assume that the heterochiasmy could be more or less attributed to the sexual selection and the sex differences in the internal environment. However, irrespectively of the proximate mechanisms, the significant sex differences in recombination rates in the Siberian jay have obvious practical implications for future work. For instance, the lower average rate of recombination in males than in females should be advantageous for QTL-mapping of genetic traits in initial low resolution analysis .
Conserved synteny, but changed marker order, in a genomic comparison with the chicken
The study confirms the remarkable degree of conserved synteny between passerines and chicken (see [4, 6–9, 13, 14]), albeit based on only a small number of comparable chromosomes. Surprisingly, loci SJ039, SJ076 and SJ101 on autosomal LGs were mapped to GgaZ, which was known to be homologous to Z chromosome in passerines [this study, [8, 9, 13]]. As argued in , this observation may suggest chromosome fusions/fissions, a spurious match, or a more complex history of the loci. Future mapping studies may help to elucidate this. Locus SJ009, which was identified within the PAR region for the Siberian jay, was found to be conserved between LGZ and GgaZ. This is consistent with previous reports of cytogenetic and genetic mapping of the Z chromosome that the PAR region was conserved between chicken and passerines [8, 88]. However, this is not the case for the other three loci (SJ039, SJ076 and SJ101) on the autosomal LGs in Siberian jay. Thus, there are more inconsistencies than consistencies in the Z-chromosomal genomic comparison between Siberian jay and chicken in this study, which is different from previous results of Z-chromosomal synteny between the passerines and chicken [9, 12, 13]. Further studies to explore the potential explanations are needed in the future. The BLAST analysis located three unlinked loci, SJ005, SJ020 and SJ034 on Gga1 and Gga2. This can be explained by the fact that these three markers had few informative meioses and therefore low power in linkage analysis and/or that these loci are located in telomeric region, which may have a higher recombination rate. Indeed, as indicated in Figure 4, these three loci are located to the end of Gga1 and Gga2 in chicken (SJ005 at 2.0 Mb of Gga1, SJ020 at 1.1 Mb of Gga1; SJ034 at 2.1 Mb of Gga2).
Although synteny was conserved, there were multiple cases of inter- and intra-chromosomal rearrangements. Similar patterns have also been observed in other passerine birds on both sex and autosomal chromosomes (see [9, 14]). As a contrast to the previous studies, the extent of chromosomal rearrangements as observed in the Gga1-LG1 and Gga2-LG2 comparisons has not previously been reported. For example, in the comparative mapping analysis between the zebra finch and the chicken, only a few instances of inversions and translocations were found for chromosome 1 . Of the total 22 homologous loci between chicken and Siberian jay found here, 12 loci (12/22, 55%) were involved in inter- or intra-chromosomal rearrangement, while a lesser proportion of the rearrangements were observed in the chicken/collared flycatcher (26/159, 16.4%)  and chicken/great reed warbler (7/44, 15.9%) comparisons . There are several possible explanations for the different rates of genome evolution, e.g. the species-specific differences in rate of mutations and/or genome evolution could have contributed to the more or less genetic similarity of different passerines to the chicken; also, it has been suggested that population structuring in species with otherwise large population sizes facilitate chromosomal rearrangements. However, the scenario outlined here should be considered with caution because (i) a small number of comparable chromosomes (n = 6) was identified here between the Siberian jay and the chicken; (ii) most of the chromosomal rearrangements are only from three chromosomes (chromosomes 1, 2 and Z). Thus, further extensive comparative mapping of genomes and genetic linkage maps of the chicken, zebra finch, great reed warbler and Siberian jay with more and denser genome coverage should provide a more detailed picture of marker order rearrangement between passerines and chicken during avian evolution. Also, more sequence data to make the genome comparison between the chicken and the zebra finch, whose genomes are only currently sequenced in birds, is worthwhile. At any rate, the important message from our comparative analyses is that a high level of observed synteny does not necessarily mean conserved marker order. If the internal rearrangements tend to occur frequently across the genome, then map information derived from one species will not be readily transferable to another.