Comparative linkage group arrangements in the more completely characterized genetic map of Arctic charr, along with comparative analyses of the linkage group and chromosome arm arrangements in rainbow trout have provided a more detailed and comprehensive understanding of the genetic map arrangements in brook charr outlined in this study. These comparative analyses indicate that there are large regions of linkage group arm retention in the genetic maps of brook charr, Arctic charr, and rainbow trout. Thirty-seven of the expected 42 linkage groups in brook charr were tentatively identified in this study. In addition, one of the singleton markers genotyped in the survey may correspond to a separate linkage group arm in rainbow trout (i.e., RT-19p), suggesting affinities to 38 possible linkage groups.
The karyotype of brook charr purportedly has 34 acrocentric and 8 metacentric chromosomes . Although the map depicted here has been assessed as having at least 8 metacentrics, up to 11 metacentrics may occur in the species, if indeed additional marker genotyping establishes that BC-15, BC-17, and BC-25 are also metacentric in structure. As evidenced from the mapping associations, it appears that two linkage groups (i.e., BC-4 and BC-16) may exhibit polymorphisms wherein some individuals may possess two acrocentrics associated with these linkage groups while others exhibit a metacentric structure. Furthermore, without a FISH analysis of the physical map, we cannot fully designate the linkage groups depicted in Table 1 as being metacentric in structure versus some type of whole-arm fusion event, similar to the karyotypic arrangements in Atlantic salmon . These linkage groups may also be representative of some partial arm translocation event. However, intraspecific chromosome rearrangements are not that uncommon in salmonids, and the alignment of distinct rainbow trout chromosome arms to the various Salvelinus linkage groups suggests that an interpretation of whole-arm rearrangements is more likely.
It is likely that the recombination rate differences identified in this study are not representative of genome-wide differences in recombination rate given the small number of comparisons used to produce these estimates. Within brook charr, comparisons were limited to 9-17 marker intervals (14-27 pairwise comparisons) and thus much of the genome was not represented in the various estimates. When multiple intervals within a single linkage group were present for comparison, recombination ratios between the parents being compared were often variable (e.g., BC-16), indicating the importance of complete genome coverage for accurate average recombination rate estimates. However, BC-16 is homologous to RT-8 in rainbow trout, and RT-8 has been reported to have extremely unusual recombination rate dynamics, in that both female and male recombination rates were observed to be greatly suppressed throughout most of the length of the linkage group . Thus, intrinsic factors regulating crossing-over mechanics may be much more variable within this particular genomic region in salmonids.
In the one case where recombination rate was significantly higher in the male mapping parent relative to the female, the loci (OMM5102/ii and BHMS465/i on BC-24) appear to be located near the telomere. Comparative mapping places OMM5102 distally (i.e., towards the telomeres) on the homologous Arctic charr (AC-3) and rainbow trout (RT-7q) linkage groups. BHMS465 has not been mapped in rainbow trout and the one copy mapped in Arctic charr maps distally on AC-24. Assuming that this pair of loci is located telomerically on BC-24, these results are not surprising in light of the work of Sakamoto et al. , who found recombination rates to be elevated in males relative to females in putative telomeric regions of the linkage group arms. Multivalent formations during Meiosis I restrict crossing over events to the telomeric regions of many chromosomes in males, thus resulting in suppressed recombination in regions proximal to the centromere and increased recombination in regions closer to the telomere [[11, 19, 20]]. In salmonids, these formations appear to be restricted to males (see  for an exception), hence the higher recombination rates observed in males relative to females in the telomeric regions of some chromosomes. Lastly, it should be noted that the higher male versus female recombination rate detected on BC-24 was not significant following Bonferroni correction.
The pairwise female: male recombination rates observed in this study among all four possible pairwise combinations of the mapping parents in brook charr (i.e., 2.41: 1) is somewhat higher than the levels observed in Arctic charr (i.e., 1.60: 1 - updated data based upon 550 map interval comparisons among the 4 mapping parents). This level of recombination is more similar to what has been observed in the rainbow trout mapping panels (i.e., ~ 2.95:1) , and much lower than levels observed in Atlantic salmon (i.e., ~ 7.23:1 - 8.26:1)[10, 21, 22]. Too few interspecific homologies existed within brook charr to permit even a preliminary analysis of whether these differences are in fact significant. A future reanalysis should permit a better understanding of this phenomenon with respect to the establishment of whether species with more acrocentric-based karyotypes do in fact have lower overall sex-specific recombination rates compared to those with more metacentric-based karyotypes, as suggested by Qumsiyeh .
Evidence exists that segregation distortion can influence marker order, estimates of map distances, and linkage relationships . While theoretical work by Hackett and Broadfoot  suggests segregation distortion at a single locus on a linkage group should have little effect on recombination estimates, the presence of two loci showing segregation distortion can result in the detection of false-positive linkage between two or more linkage groups . However, these models are based upon tests of zygotic segregation distortion resulting from tests of combined parental genotypic combinations, such as those implemented when trying to build consensus genetic maps. Tests of gametic segregation distortion (conducted in this study) for assessing sex-specific genetic maps are expected to have less pronounced effects on linkage map construction and are a more accurate method of assessing such differences . Even gametic phase distortion may, however, be associated with increased estimates of recombination distances between linked markers . Localization of markers around recombination 'hot-spots' may also lead to a disruption in marker orders, but this effect may only be pronounced in regions of the salmonid genome involved with quadrivalent formations during meiosis (e.g., male meioses) .
All mapping parents except the LN4F contained markers on BC-16 which exhibited significant segregation distortion prior to Bonferroni correction. This might have partially accounted for the variability in map distances observed among mapping parents. Interestingly, the homologous linkage group to BC-16 in rainbow trout (i.e., RT-8) has a large degree of recombination suppression in females [11, 12]. This region is also of interest evolutionarily as RT-8 appears to contain one or more genes important for several life-history traits, including development rate [27, 28], spawning time [29 - 31] and maturation timing . It has been argued that reduced recombination can be adaptive in that it helps to preserve highly compatible combinations of genes or gene complexes . Whether the high degree of segregation distortion observed for markers on BC-16 stem directly from the importance of the genes within this linkage group region is unclear (i.e., are slight genomic incompatibilities more pronounced within BC-16 due to disruption of co-adapted gene complexes?). The mapping of additional markers to BC-16 within these experimental mapping panels and indeed additional brook charr families would assist in understanding the recombination 'hot-spot' dynamics within this chromosomal region.
Lastly, it should be noted that segregation distortion rates might be elevated in the HL brook charr due to their hybrid history. In Arctic charr, segregation distortion is elevated in hybrid relative to pure strain families . Given the history of hybridization in the HL strain brook charr , the frequency of segregation distortion in HL brook charr might be elevated relative to that of pure strain brook charr. With so few markers currently mapped in LN brook charr, even comparing relative frequencies of segregation distortion between HL and LN brook charr is not particularly informative and thus this relationship cannot be tested at present. Therefore, it is important to recognize that segregation distortion rates observed in Hill's Lake brook charr might be an overestimation of typical rates for pure strain brook charr.
The observation that SEX is unlinked to the SSOSL32 marker in brook charr but appears tightly coupled to the Ots500NWFSC marker on BC-4 is intriguing given the polymorphisms with SEX linkage reported in Arctic charr . In Arctic charr, SSOSL32 and Ots500NWFSC variation has been consistently linked to SEX, while other markers on AC-4 have shown variable associations. This could be due to the fact that this linkage group may be split into two acrocentric arms in certain individual males and yet be retained as a metacentric chromosome in other males, or result from pseudolinkage . A similar polymorphism appears to exist in brook charr. Given that SSOSL32 and Ots500NWFSC appear to map close to one another in the central portion of the AC-4 linkage group , it is possible that some type of inversion has occurred in the homologous region of BC-4 to uncouple this association. This may have resulted in the placement of SSOSL32 onto the linkage group arm that shows variable associations with SEX in Salvelinus. Clearly, the examination of this sex linkage association in additional families of brook charr is needed. Ideally, this should be conducted across multiple strains of the species. In addition, a more complete genotyping survey of the markers located on this linkage group is needed in order to assess more precisely the 'break-points' in the SEX: marker associations, and define more accurately the recombination distances between SSOSL32 and Ots500NWFSC in brook charr.
Homeologous and homologous affinities
Given that the modal number of chromosome arms in salmonids is 100  and most Actinopterygiians have diploid chromosome numbers of 48 or 50 [35, 36], it is expected that up to 25 homeologous affinities would be present in salmonids. As brook charr still possess the expected number of doubled chromosome arms following polyploidization, it was expected that any one brook charr linkage group would show homeology to only one other brook charr linkage group if the linkage group was representative of an acrocentric chromosome, or at most two other brook charr linkage groups if representative of a metacentric chromosome. Although greater than 25 putative homeologous affinities have been detected in the current study the expectation of 1:1 arm homeologies were largely met suggesting that there is a propensity to largely maintain evolutionary linkage arm arrangements in the salmonids. No homeologies were observed for five linkage groups (i.e., AC/BC-31, -33, -36, -39, and -43), in the combined Salvelinus linkage maps. Markers from AC/BC-31 are only syntenic with those on RT-16q (see Additional File 8, 9) confirming the status of this linkage group. For the other 4 linkage groups, it is possible that the designated linkage groups are only part of a larger linkage group that has not yet been identified (i.e., lack of intercalary markers genotyped to join the separated clusters), given that markers on each of these linkage groups assign to 2 - 4 different rainbow trout linkage group arms. However, for markers on AC/BC-36, -39, and -43, two or more markers define their assignments to rainbow trout linkage group arms RT-5p, -19q, and -9p. In each instance, these are the major cross homologies to the rainbow trout map suggesting that these are valid single linkage group arms. For AC/BC-33, assignments are possible to RT-11 (based upon a single marker homology), but this acrocentric linkage group also shares homology to AC/BC-4, -22, and therefore, further research is required to define this relationship. Since most assignments of cross homology for RT-11 are to AC/BC-22, it is possible that AC/BC-33 represents an unlinked fragment of AC/BC-22.
Linkage groups AC-15/BC-15 would appear to be a metacentric linkage group in structure, and along with AC-1/BC-1 and AC-3/BC-3, are three metacentric linkage groups that appear to have maintained a conserved structure between Salvelinus and Oncorhynchus. AC/BC-1, -3, and -15, are homologous to metacentric groups RT-29, -15, and -10, respectively, in rainbow trout. Only 1 arm of AC/BC-15 has been identified as possessing a homeologous affinity to AC/BC-10, and this pair of arms corresponds to the RT-10/18 homeology grouping . The arm from AC-15/BC-15 lacking homeologous affinities to AC-10/BC-10, is homologous to RT-10p arm (see Additional File 9), supporting the contention that AC-15/BC-15 represent metacentric chromosomes.
Four of the duplicated genetic markers genotyped had only a single copy assigned to a known linkage group, while the other duplicate is currently recorded as a singleton. Unlinked copies of BX073647 and CA368462 appear to be homologous to rainbow trout linkage group arms where there is no coverage by the current genetic map for brook charr. BX073647 maps to homeologous linkage groups RT-17p/22p, and CA368462 maps to homeologous linkage groups AS-17/33, which is also homologous to the RT-17p/22p linkage group pair in rainbow trout . Cross homology assignments would suggest that BX073647 and CA368462 are located on AC/BC-8/18 linkage groups, which needs to be confirmed with additional marker genotyping.
Four putative homeologies (BC-6/35, BC-14/30, BC-16/37 and BC-27/35) identified in brook charr, have not been detected to date in Arctic charr. While most of the identified homeologies in brook charr appear to represent conserved, known homeologous affinities in Arctic charr, rainbow trout, and Atlantic salmon, two of these (i.e., BC-14/30 and BC-16/37) appear to represent relationships which are not currently identified in other salmonine species. BC-14 shows homology to RT-19p (single marker) and RT-24q (multiple markers), while BC-30 shares homology to RT-3p and RT-6p (each with single marker affinities). Regions on RT-6p, and 19p are derived from the M ancestral karyotypic lineage in teleost fishes , suggesting that this homeology may be representative of either a 3R or 4R WGD homeology. BC-16 shares homology to RT-8 while BC-37 is related to RT-21p. There is a small segment on RT-21p derived from the F ancestral lineage of fishes, while it appears that most of the RT-8q arm is derived from the M ancestral lineage, and the RT-8p arm from the I lineage. Current data therefore, do not reconcile an origin for this homeology (i.e., BC-16/37) from the known duplicated segments in salmonids. With respect to the BC-6/35 and BC-27/35 homeologies, the former grouping shares affinity to the RT-27p/31p duplications (= B ancestral lineages), while the BC-27/35 region may relate to RT-6p/27q duplications (= possible K ancestral lineages), although it should be mentioned that the ancestral origins for the RT-6p arm are not well established , and therefore the assignment to other ancestral groupings may be revealed.
Interestingly, the majority of duplicated markers identified in brook charr correspond to linkage groups in rainbow trout and Atlantic salmon where the highest number of duplicated markers have been detected. Eight of the 12 conserved homologies between the duplicated homeologs in the two charr species and rainbow trout (Sal-23/35 and RT-27/31, Sal-3/24 and RT-7/15, Sal-12/27 and RT-12/16, Sal-1/21 and RT-2/29, BC-13b/34, AC-13/34 and RT-14/20, Sal-10/15 and RT-10/18, and BC-20a/20b and RT-2/9) are supported by a high number of duplicated markers in rainbow trout [10, 13]. In addition, three of the six homeologies conserved between charr species and Atlantic salmon (BC-1/21 and AS-1/6, BC-13b/34 and AS-19/28, and BC-20a/20b and AS-4/11) are Atlantic salmon homeologies currently supported by the highest number of duplicated markers . Phillips et al.  also observed a high degree of correlation in the number of duplicated markers supporting homeologous associations in both rainbow trout and Atlantic salmon. The observation that several of these markers remain duplicated in brook charr provides additional evidence for the continued exchange of information between these homeologous linkage groups across the Salmoninae. It remains unclear, however, why these chromosomal regions in particular contain such a high frequency of conserved duplicated markers.
The regions possessing the highest retention of duplicated markers are also those regions most likely to exhibit pseudolinkage in these species. These regions were homologous to RT-2p/29q; RT-2q/9q; RT-7q/15p; RT-12p/16p; RT-27p/31p in rainbow trout . Here we report the expression of a pseudolinkage region on BC-1/21 (= RT-2p/29q homology). Apparent pseudolinkage arrangements have also been detected in the male mapping parents in two Arctic charr mapping panels involving three different linkage group regions (i.e., AC-12/27 (= RT-12p/16p); AC-6/23 (= RT-27p/31p); and AC-4/25a (= RT-11/12q/26)) [15; current study]. Hence, there appears to be a high degree of retention in the propensity to form quadrivalent pseudolinkage arrangements within a specific subset of the salmonid genome.
The suppression of diploidization in these linkage groups due to the continued exchange of chromosomal segments (especially in the telomeric regions of male quadrivalent formations) would ensure the continued retention of duplicated marker expression in populations exhibiting such phenomena. These meiotic processes would also shelter genetic markers towards the central parts of male metacentric linkage groups from recombination during meiosis . Thus, it is tempting to speculate that one of the evolutionary forces driving these unusual meiotic processes in male salmonids is selection for co-adapted gene complexes. Whether synteny blocks prove to be less re-arranged within the centromeric regions of linkage groups exhibiting pseudolinkage regions awaits the completion of greater genomic sequence data.