INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

The modern domestic dog features more than 300 well-defined isolated breeding populations, most of which are readily accessible for mapping genes involved in complex phenotypes (Ostrander and Giniger 1997; Galibert et al. 1998; Ostrander et al. 2000). Indeed, extensive canine pedigrees, coupled with an enormous diversity in morphology and behavior, provides a valuable mechanism for understanding the genetic regulation of mammalian growth and development. In addition, the mating of closely related individuals in order to maximize segregation of desirable traits has led to the propagation of autosomal recessive diseases in the modern dog, many of which are similar or, in some cases, identical to human diseases (e.g., Dodds 1989; Menon et al. 1992; Sharp et al. 1992; Stolzfus et al. 1992; Henthorn et al. 1994; Zheng et al. 1994; Acland et al. 1998, 1999; Lin et al. 1999; Veske et al. 1999). To date, ~360 genetic diseases have been identified in the dog, representing the largest number reported for any nonhuman mammal (Patterson 2000, 2001). Genetic linkage analysis utilizing domestic dog pedigrees, therefore, should provide a unique and efficient mechanism for understanding the molecular and cellular biology of human health and disease (Patterson et al. 1988; Ostrander et al. 2000).

In the past three years, we and our collaborators have made major advances in the production of a canine meiotic linkage map (Lingaas et al. 1997; Mellersh et al. 1997; Neff et al. 1999; Werner et al. 1999) and a radiation hybrid (RH) map (Priat et al. 1998; Mellersh et al. 2000). The recently published meiotic linkage map was composed of ~350 markers organized into linkage groups for 37 autosomes and the X chromosome (Werner et al. 1999), whereas the RH map was composed of 600 markers, including 218 genes and 382 microsatellites, organized into 77 RH groups (Priat et al. 1998; Mellersh et al. 2000). By taking advantage of 216 markers mapped on both a set of three-generation reference families (Mellersh et al. 1997) and on a well-characterized canine RH panel (Vignaux et al. 1999), integration of the linkage and RH maps has been accomplished, leading to the production of a 724-marker map (235 genes and 489 microsatellites) with a marker spaced, on average, every 3.7 cM or 24.3 cR₅₀₀₀ (Mellersh et al. 2000). The evolutionary relationship between the canine and human genomes is also now well established through the efforts of two independent groups, both of whom have used bidirectional heterologous chromosome painting to identify evolutionarily conserved regions of human chromosomes (reciprocal Zoo-FISH; Breen et al. 1999c; Yang et al. 1999). Specifically, Breen et al. (1999c) and Yang et al. (1999) propose that 68 and 73 evolutionarily conserved segments, respectively, from the human genome can be visualized by reciprocal Zoo-FISH. Most recently, Sargan et al. (2000) expanded the latter of these reports and assigned all linkage, RH, and syntenic groups from the most recently published canine genome map (Mellersh et al. 2000) to specific dog chromosomes.

Although the accuracy and utility of the map are evidenced by the localization of disease loci for canine renal cancer, narcolepsy, multiple forms of retinitis pigmentosa, hematologic disorders, and several metabolic diseases (e.g., Henthorn et al. 1994; Yuzbasiyan-Gurkan et al. 1997; Acland et al. 1998; Lingaas et al. 1998; Lin et al. 1999; Jónasdóttir et al. 2000), at this juncture those wishing to exploit the dog system for refined mapping and cloning of complex diseases will be met with several significant obstacles. The placement of a marker every 3-4 cM is sufficient to ensure that genome scans of canine families will identify linked markers, but a denser map of more highly informative microsatellites is clearly needed for refining regions defined by linkage, as well as to facilitate the mapping of disease loci within highly inbred families. A more precise understanding of the evolutionary relationship between the canine and human genomes is needed in order that all evolutionary breakpoints may be mapped, and data from the more comprehensively mapped human and murine genomes used for the selection of candidate genes in the study of diseases. The meiotic linkage and RH maps must be fully integrated with the cytogenetic map. The latter represents a particularly challenging task because the similarity in size and banding patterns of many dog chromosomes renders them notoriously difficult to identify by classical cytogenetic methods. Using conventional cytogenetics, the International Committee for the Standardization of the Karyotype of the Dog was able to standardize the karyotype for dog chromosomes 1-21 and the sex chromosomes (Switonski et al. 1996). The Committee then assigned the chromosome-specific paint probes developed by Langford et al. (1996) and recommended that the chromosome numbering of Reimann et al. (1996) should be used for the remaining autosomes (Breen et al. 1998, 1999a). In this report we describe our collective effort to produce an integrated map of the dog genome incorporating cytogenetic, RH, and meiotic data. Building upon a newly developed set of 266 FISH-mapped chromosome-specific cosmid clones, each containing a microsatellite marker that unambiguously identifies one of the 38 autosomes of the dog, we are able for the first time to assign all dog chromosomes to their corresponding meiotic linkage and RH groups. The resulting map features 1078 microsatellites, 320 gene-based markers, and a total of 302 chromosome-specific markers, all of which are assigned by FISH using what is now agreed upon as the standardized chromosome nomenclature. This more than doubles the previously reported density of markers. All known evolutionary breakpoints in the human map are assigned to the canine cytogenetic map and therefore to the RH and linkage maps. Finally, RH mapping of proximal and/or distal markers that were initially mapped by FISH has allowed all but four integrated groups to be orientated. The resulting map, which features robust anchor points on every dog chromosome and a high density of polymorphic markers, provides an invaluable resource for the mapping and cloning of canine genes of interest.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

In this paper we present an integrated cytogenetic/RH and meiotic map of the dog. We localized 266 microsatellite-containing cosmid clones on all dog chromosomes through FISH analysis, assigning 1500 markers by RH analysis and 354 markers by meiotic linkage analysis. Through sets of common markers, 251 RH/genetic, 102 RH/cytogenetic, and 52 linkage/cytogenetic, one or more RH groups and one genetic linkage group were assigned to each dog chromosome. This represents an integrated map with 1800 unique markers. In total, 72 RH groups are assigned to all chromosomes of the dog and 39 meiotic linkage groups are assigned to all autosomes plus the X chromosome. The resulting map (poster enclosed with this issue as Fig. 1) covers >90% of the genome.

Cytogenetic Map

Chromosome-Specific FISH Probe Selection

RH Map: Marker Characteristics

Microsatellite Markers

Gene-Based Markers

Chromosome-Specific Markers

PDE

Radiation Hybrid Mapping

The 900 new markers added here, as compared with the map reported by Mellersh et al. (2000), were typed on the 118-cell-line RHDF5000-2 panel (Vignaux et al. 1999). Data from these markers were added to the previous 600-marker RH map (Mellersh et al. 2000) and computed using the MultiMap software. Pairwise linkage analysis at LOD > 8.0 resulted in 89 RH groups. A total of 1413 out of 1500 markers were incorporated into these RH groups, with the remaining 87 markers remaining unlinked. The presence among RH markers of 102 chromosome-specific markers allowed 72 of the RH groups to be linked to one another resulting in 40 larger groups, each of which was assigned to a specific dog chromosome (Fig. 1). The 17 remaining orphan RH groups, which include 2-12 markers, have a cumulated size of 586 cR₅₀₀₀, representing only 2.5% of the map (data not shown).

The marker order for each RH group was obtained using Multipoint analysis. The order was initially determined using an LOD score of >3.0 (likelihood odds = 1000/1), resulting in a framework map consisting of 579 markers (45%). These markers are underlined in Figure 1. All remaining markers were then integrated relative to the framework markers to give the comprehensive map presented here. For RH groups presenting inconsistent ordering data, the initial pairwise analysis was repeated using a higher LOD score value (LOD > 9) that resulted in the formation of more RH groups. Ordering could then be carried out by multipoint analysis. The number of RH markers assigned to each autosome ranged from 67 markers (CFA 1) to a minimum of 8 (CFA 38), with the smallest dog chromosome, Y, having only 4 markers. The size of RH groups assigned to each chromosome ranged from 1272 cR₅₀₀₀ (CFA 1) to 112 cR₅₀₀₀ (CFA 38), with 17 cR₅₀₀₀ for the Y chromosome. The total size of the RH map is 23,428 cR₅₀₀₀ with 1413 markers mapped to 1354 unique positions. Thus, the average intermarker distance is 17 cR₅₀₀₀. Compared with the previously reported RH map (Mellersh et al. 2000), the present map represents a 2.5-fold increase in marker density, and decreases the average intermarker distance 1.5-fold.

Using maximum likelihood predictions, the canine genome has been estimated to be 26.5 M (Neff et al. 1999). The total length of the RH map presented here is 23,428 cR₅₀₀₀. Taking into account the cR/kb correspondence (see below) and the correspondence of 1 Mb/1 cM, the total size of the RH map thus amounts to 23.5 M. This map, therefore, is estimated to cover 88% of the canine genome. In agreement with this coverage estimate, only 7% of the markers remain unlinked at this point of the RH-map construction.

Meiotic Map: Marker Characteristics

Chromosome-Specific Microsatellite Markers

Meiotic Linkage Mapping

Microsatellites from each of the 38 dog autosomes were genotyped on dog reference families that have been described previously (Mellersh et al. 1997), and the data were merged with those from the previous meiotic linkage map described by Werner et al. (1999). We identified 39 linkage groups, containing a total of 350 markers. Of these 38 groups could be assigned to a chromosome via a chromosome-specific microsatellite marker. An additional linkage group did not contain a chromosome-specific microsatellite and could not, therefore, be conclusively assigned to a chromosome (data not shown).

The chromosome-specific microsatellites could be assigned to their respective linkage groups with varying levels of significance. For 32 autosomes, CFA 1-12, 14-27, 29, 31, 33, 35-37, the chromosome-specific microsatellites were linked to other microsatellites in the corresponding group with LOD scores >5.0, indicating significant evidence of linkage. Meiotic linkage data for the six remaining chromosomes were as follows: the chromosome-specific microsatellite for CFA 13 (AHTH310) was linked to one other marker (FH2394) in the group with a LOD score >4.0, and to multiple other markers in the same group with LOD scores >3.0. The markers representing CFA 28 (AHTK135), CFA 32 (AHTH327), and CFA 34 (AHTH163) were each linked to a single marker (LEI006, CPH2, and FH2377, respectively) in their corresponding groups with a LOD score >3.0. For CFA 30 (LEI-1F11) and CFA 38 (AHTH91), the best two-point LOD scores were 2.709 (to PEZ7) and 2.912 (to FH2244), respectively. However, RH data for these two markers did enable them to be assigned to their respective groups with significant LOD scores (>8.0), thus validating the assignment of the whole group.

The Integrated Map: Integration of the Three Maps

The use of FISH analysis to integrate the RH and meiotic maps with the cytogenetic map is illustrated in Figure 2. A comparison of the RH, meiotic, and cytogenetic maps shows that they are highly concordant. The only observed discrepancies in the colinearity between the maps is the inversion of a segment containing four markers on CFA 4, and 20 inverted pairs of markers distributed on other autosomes. Chromosome-specific FISH-mapped markers were RH-mapped on all chromosomes as follows: 8 dog chromosomes harbor 3-6 probes, 20 have 2 probes, and 12 chromosomes have 1 probe. RH/meiotic groups could be readily orientated on 33 chromosomes owing to the localization of at least one FISH probe at one end of the chromosome. Another chromosome, CFA 19, could be oriented because of syntenic conservation between human and dog chromosomes. The maps for the remaining four autosomes (i.e., CFA 21, CFA 32, CFA 35, and CFA 36), each harboring one midchromosome probe, and for the two sex chromosomes cannot as yet be orientated (Fig. 1).

View larger version (37K): [in this window] [in a new window]   Figure 2   Multicolor fluorescence in situ hybridization of canine clones to dog chromosome 5 (CFA 5). The integrated RH and meiotic maps for CFA 5 are shown on the right hand side with their corresponding sizes noted below each map. Six clones were selected on the basis of their positions along the length of the RH map and are indicated in colored text. A seventh clone, SLC2A4, was selected on the basis of its position in the meiotic map. Each of the seven clones was labeled with one of the following five fluorochromes: Spectrum Green dUTP (green signal), Spectrum Orange dUTP (gold signal), Spectrum Red dUTP (red signal), DEAC (aqua signal), and Cy5 (pink signal). All seven clones were cohybridized by FISH to elongated dog chromosomes. The resulting FISH image of a DAPI-counterstained CFA 5 is shown in the middle of the figure, illustrating the hybridization signals of all seven clones, along with the assignment of each clone to the DAPI-banded ideogram of CFA 5 (Breen et al 1999a). On the far left is the enhanced DAPI-banded image of the same CFA 5 alongside the corresponding conserved human chromosome segments identified by Breen et al. (1999c).

By comparing the size in megabases of a chromosome (deduced by the bivariate flow-sorting measurements of Langford et al. 1996) to the size of the RH group harboring FISH markers located at the two ends of the chromosome (this was possible for CFA 1, CFA 10, CFA 12, CFA 14, and CFA 20), a correspondence of 1 cR₅₀₀₀ for 100 kb was calculated. This result compares well with those reported for human and murine genomes at comparable radiation doses (McCarthy et al. 1997; Stewart et al. 1997).

Coverage of the Integrated Map

Our data indicate that RH groups span the length of all but three autosomes, CFA 32, CFA 36, and CFA 38, for which we observed a clear difference between the size of the chromosome and that of the corresponding RH group. Specifically, CFA 32 and CFA 36 are each composed of one RH group, covering approximately half of the chromosome, whereas CFA 38 is covered by one RH group representing approximately one-third of the chromosome. The X chromosome is more poorly covered by two RH groups and four single markers. Regions not covered by the RH map can be estimated to 200 Mb, or ~8% of the total genome size. However, unlinked markers and orphan RH groups are likely to fill the gaps.

The size of meiotic linkage groups in Figure 1 approximates to their chromosome length for all but eight autosomes: CFA 6, CFA 17, CFA 19, CFA 21, CFA 24, CFA 32, CFA 34, and CFA 36. Coverage of chromosomes by their corresponding meiotic linkage group is 60% for CFA 19 and CFA 21, 50% for CFA 6 and CFA 24, 30% for CFA 17 and CFA 32, 20% for CFA 36, and 10% for CFA 34. The X chromosome is covered at 50%. However, by combining data from the RH and meiotic linkage maps, only half of both CFA 32 and CFA X and two-thirds of CFA 36 are not represented in the integrated map. A summary of the relevant statistics of the integrated map is shown in Table 3.

Comparative Mapping

For each dog chromosome presented in Figure 1, we have illustrated the regions of evolutionarily conserved chromosome segments that were observed in the Zoo-FISH studies of Breen et al. (1999c). Although Yang et al. (1999) used a different chromosome nomenclature from that recently endorsed (International Society for Animal Genetics 2000, G. Dolf, DogMap Chairman, pers. comm.) and in use here, correspondence between the two chromosome numbering systems was tentatively identified by Sargan et al. (2000). Our localization of 320 gene markers on the RH map has now verified this correspondence. This has allowed us to confirm that these two studies were grossly comparable in their assessment of the distribution of human conserved chromosome segments throughout the dog karyotype. We observed differences in five cases. Four differences were at the proximal end of CFA 4, CFA 18, and CFA 31 and the telomeric end of CFA 25. This was because of the lack of visible signal at the centromeric and telomeric ends of the dog chromosomes reported by Breen et al. (1999c), a phenomenon that has also been reported for other species (Jauch et al. 1992; Wienberg et al. 1992; Nash et al. 1998). However, hybridization signals were observed in these regions by Yang et al. (1999), and we have now confirmed the presence of these regions by identification of corresponding type I markers in the RH map. The additional information for these four regions provided by Yang et al. (1999) is therefore included in Figure 1, highlighted with (Y) beside the data. A fifth difference was observed toward the distal end of CFA 10, where Yang et al. (1999) reported a small block of HSA 12 between blocks of HSA 22 and HSA 2 reported by Breen et al. (1999c). This finding has not yet been confirmed by RH mapping. Zoo-FISH data between dog and human identified no fewer than 67 (Breen et al. 1999c) or 75 conserved chromosome segments (Yang et al. 1999). RH mapping of dog genes whose human locations are known has thus far allowed the detection of 65 conserved segments.

For 229 RH-mapped dog genes, the human localization of the putative ortholog was known. All but two dog chromosomes (CFA 32 and CFA 36) harbor at least one (and up to 22) gene marker for which a physically mapped human ortholog is known. The absence of any mapped gene on CFA 32 and CFA 36 precluded the use of the RH data for comparative assessment of these two autosomes. Conserved syntenic fragments between dog and human are indicated in Figure 1, with colored boxes to the immediate left of the RH markers. Apparently 18 dog chromosomes correspond to only one human fragment each, that is, CFA 8/HSA 14q, CFA 12/HSA 6pq, CFA 14/HSA 7pq, CFA 21/HSA 11pq, CFA 22/HSA 13q, CFA 23/HSA 3pq, CFA 24/HSA 20pq, CFA 27/HSA 12pq, CFA 28/HSA 10q, CFA 29/HSA 8q, CFA 30/HSA 15q, CFA 33/HSA 3q, CFA 34/HSA 3q, CFA 35/HSA 6p, CFA 37/HSA 2q, CFA 38/HSA 1q, X/X, and Y/Y. Another 13 dog chromosomes (CFA 7, CFA 9, CFA 10, CFA 11, CFA 13, CFA 15, CFA 16, CFA 17, CFA 18, CFA 19, CFA 20, CFA 26, and CFA 31) correspond to two human fragments each, and the remaining 9 dog chromosomes correspond to three or four different human chromosomal fragments each. In contrast, only four human chromosomes share exclusive conserved synteny with a dog chromosome (HSA 14/CFA 8, HSA 20/CFA 24, HSA 21/CFA 31, and chromosome X). The remainder are split into two to eight chromosomal segments (e.g., HSA 1 is split into eight fragments in the dog, corresponding to regions of CFA 2, CFA 4, CFA 5, CFA 6, CFA 7, CFA 15, CFA 17, and CFA 38; Breen et al. 1999c; Yang et al. 1999).

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

Great strides have been made over the past few years toward the development of a canine genome map. Prior to this study, however, there were a number of limitations associated with the canine map. Chromosomal assignment of meiotic linkage groups had been made for only 19 of the 38 dog autosomes (Mellersh et al. 2000), with the remainder being referred to as anonymous linkage groups. In addition, the chromosomal orientation of the meiotic linkage and RH groups was largely unknown. Furthermore, estimates of genome coverage had been entirely theoretical, and little was known about the extent to which the RH and meiotic linkage maps extended along the length of their respective chromosomes. In this study we have addressed these issues. We have conclusively assigned 266 cosmid clones to their specific dog chromosomes, by FISH, using the chromosome numbering system endorsed by the ISAG 2000 DogMap workshop (G. Dolf, DogMap Chairman, pers. comm.). Because all the cosmid clones had been prescreened to contain a microsatellite repeat, they represent not only additional chromosome-specific genetic markers but also the key resource that enabled us, for the first time, to integrate the canine meiotic linkage and RH maps with the cytogenetic map.

Meiotic linkage data from 38 chromosome-specific microsatellite markers, derived from these clones, were used to anchor meiotic linkage groups to each of the 38 autosomes. RH mapping of 102 chromosome-specific markers similarly allowed us to assign 72 RH groups to all chromosomes comprising the dog karyotype, and provide an integrated cytogenetic, meiotic linkage, and RH map of the dog. Moreover, FISH mapping of other previously RH-mapped markers strengthen the chromosomal assignments made in this study, for example, GALK1, RARA, THRA1, NF1, and KRT17 on CFA 9 (Werner et al. 1997; Miller et al. 1999), C04107 on CFA10 (Van der Sluis et al. 1999), PROP1 on CFA11 (Lantinga-van Leuween et al. 2000a), and PIT1 on CFA31 (Lantinga-van Leuween et al. 2000b). Sixteen of the 19 integrated assignments reported by Werner et al. (1999) and Mellersh et al. (2000) are in agreement with the nomenclature used in this study. However, there are three discrepancies in chromosome nomenclature. In this report the chromosomes we refer to as CFA 17, CFA 23, and CFA 27 were described by Werner et al. (1999) and Mellersh et al. (2000) as CFA 15, CFA 22, and CFA 16, respectively. Reciprocal exchange of probes from both laboratories now confirms that the nomenclature used here is correct (P. Werner and M. Breen, data not shown). Also in this study, when possible, we used RH and meiotic linkage mapping to position chromosome-specific markers that had been FISH-mapped to both the proximal and distal ends of each chromosome. This approach allowed us to orientate all but four of the integrated RH and meiotic groups on the chromosomes to which they have been assigned. The exceptions are those groups assigned to CFA 21, CFA 32, CFA 35, and CFA 36, which will require additional markers to reveal their true orientations. In the previous meiotic (Werner et al. 1999) and RH (Mellersh et al. 2000) maps, calculation of the coverage was based on theoretical considerations. Integration of the RH, meiotic, and cytogenetic maps allowed a more accurate assessment of chromosome coverage, as is demonstrated clearly in Figure 1. Good coverage is evident for all autosomes, with the exception of half of CFA 32 and X as well as two-thirds of CFA 36. The integrated RH/cytogenetic map would, therefore, confirm genome coverage >90%. Chromosomes for which the map coverage is less than complete can now be targeted by specific efforts to provide full coverage. For instance, the meiotic linkage maps for three chromosomes in particular, CFA 6, CFA 19, and CFA 24, appear to be short (to varying extents) at the proximal and/or distal ends, and the map for CFA 17 appears not to represent the distal half of the chromosome. In all these cases we will aim to select the markers at the ends of both the RH and meiotic maps, in order to isolate BAC clones that may then be FISH-mapped to determine the true extent of the chromosome coverage of the existing maps.

The meiotic linkage groups assigned to seven chromosomes (CFA 21, CFA 32, CFA 33, CFA 34, CFA 35, CFA 36, and CFA 38) do not yet contain a sufficient number of appropriately placed markers to accurately determine the extent of chromosome coverage. We are therefore pursuing these chromosomes from a comparative angle by targeting type I markers that have been mapped in the corresponding evolutionarily conserved chromosome segments as identified by Breen et al. (1999c) and Yang et al. (1999).

The present study provides a significant advance in the development of an effective genome map of the dog. The genome-wide integration of the meiotic and RH maps with the cytogenetic map, together with the increase in the number and density of RH-mapped markers, will improve our ability to use the canine genome to identify important genes in major ways. It will now be possible to assign accurately markers linked to canine disease genes to their corresponding position in the human genome, using the set of conserved genes presented in this study in combination with previously established gross levels of conserved synteny, reported by Breen et al. (1999c) and Yang et al. (1999). This, in turn, will provide a mechanism for identifying appropriate positional candidate genes whose disease-causing roles can subsequently be investigated. In addition, the significant increase in the number of informative microsatellites on the RH map will facilitate the direct identification of a disease-associated gene from that of a linked marker. The 561 polymorphic di- and tetranucleotide microsatellites in the integrated map will provide complete coverage of the dog genome because any position in the genome is within an average of 10 cM of at least one of these microsatellites. Finally, this integrated map will facilitate the cloning of desired genes, because it allows us to exploit the dog as a powerful new genetic system for mapping complex traits that have been difficult to identify through corresponding studies of human families.

	METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

Generation of Chromosome-Specific Data

A dog cosmid genomic library in pWE15 (Stratagene) was gridded manually and screened for the presence of CA/GT repeats, using high stringency, as described previously (Holmes et al. 1993). A total of 2304 cosmids were screened, and 328 (26%) were found to be positive. DNA from all positive clones was prepared using either a QIAprep (QIAGEN) or Wizard Plus SV (Promega).

Fluorescence in Situ Hybridization (FISH)

Image Acquisition and Chromosome Assignment

RH Map Data

Generation of RH Markers

Chromosome-Specific Markers

RH Typing and RH Map Construction

Meiotic Linkage Data

A subset of 38 chromosome-specific cosmid clones was processed to characterize their respective microsatellite markers.

Subcloning and Identification of Microsatellite Markers

Design of Suitable Locus-Specific PCR Primers

Genotyping of Chromosome-Specific Microsatellites

Meiotic Linkage Map Construction

Human/Dog Comparative Mapping

Orthologous human genes have been defined by BLAST searches (Altschul et al. 1990) against public databases (GenBank "nr" and "HTGS") in February 2001; default BLAST criteria were used. Chromosomal locations have been found in GeneAtlas (http://www.citi2.fr/GENATLAS) and LocusLink (http://www.ncbi.nlm.nih.gov/LocusLink).

Web Sites

Comprehensive data of the integrated map (tables and figures) and the characteristics of all markers will appear on the following linked Web sites: http://www-recomgen.univ-rennes1.fr/doggy.html, http://www.aht.org.uk/cytogenetic-map/dog.html, and http://www.fhcrc.org/science/dog_genome/dog.html.