Comparative radiation hybrid mapping

Development of a bovine whole genome radiation hybrid map for comparative mapping
across species and the identification of positional candidate genes for genetically mapped
traits

Since 1993, a major international effort has been put into the creation of genetic maps of the bovine genome - there are now over 1000 loci mapped (Barendse et al 1994, 1997, Bishop et al 1994, Georges et al 1995) - while recent efforts have integrated the bovine genetic and physical maps (Ferretti et al 1997). The number of markers mapped is now at or beyond the resolving limit of 3-5cM that can be achieved with the reference populations used. Nearly 80% of genetically mapped loci have been anonymous microsatellites. Whilst it would be desirable to have a greater number of expressed sequences (ESTS) on the genetic map, as candidate genes for traits and to align comparative maps across species, generally they have limited polymorphism which makes their large scale genetic mapping an unrealistic option. Thus another approach is required to map ESTS at high density.

Trait mapping

The genetic maps of cattle are currently being used to identify trait loci through segregation analysis. To date six monogenic traits and five putative QTL have been localised within 20cM intervals on bovine chromosomes ( Georges et al 1993, Charlier et al 1995, Georges et al 1995). In addition, there are several large trait mapping studies underway and it is likely that over the next few years many QTL will be localised to chromosomal regions. Identification of the genes involved in a trait from the map positions is itself a major task. The most attractive approach to identifying the trait genes is candidate positional cloning, whereby the region in which the trait has been mapped is examined for genes which are likely to be involved in the trait. These candidate genes are then tested for a role in the trait. For this approach to be successful it is important that genes, or ESTS, are mapped at high density, or that the alignment of the bovine map with those of other species, such as man which is gene rich, is detailed enough to allow accurate transfer of information.

The recent identification of myostatin as the gene responsible for the double muscle phenotype (mh) in Belgian Blue and other breeds of cattle was achieved through a positional candidate gene approach (Grobet et al 1997). Myostatin was first identified as being responsible for muscular
hypertrophy in mice, this gene had not been mapped in man so Grobet et al transferred map information from mouse to man using a human radiation hybrid panel, which placed myostatin next to a collagen gene. This collagen gene had been mapped in cattle and was located close to the linkage mapping position for mh. Sequencing of myostatin from mh and non-mh cows showed that an 11 bp deletion was responsible for the mh phenotype.

At present, the alignment of maps between species is at low resolution and the presence of the appropriate gene on bovine chromosome 2 to infer the myostatin location from the human map position was fortuitous. Availability of a radiation hybrid map for cattle will facilitate the direct and reliable transfer of information from mouse or man to cattle.

Comparative mapping

Recently there has been a major effort to map ESTS in man
eg. the IMAGE consortium has sequenced over 500,000 cDNA clones and assigned 10,000 ESTS to the human map. It is clearly important to be able to use this information from man in searching for candidate genes in cattle. The conservation of regions of chromosomes between cattle and other species, in terms of genes present, has been demonstrated though the use of somatic cell hybrid panels and heterologous probes (eg. Womack et al 1995) and through chromosome painting (eg. Solinas-Toldo et al 1995, Hayes et al 1995). However, it is not known how this conservation is maintained at the level of gene order, or the precise location of break points, as at present, knowledge of conservation of synteny is relatively crude with information on relatively few genetically mapped loci in common between species.

Radiation hybrid mapping

Whole genome-radiation hybrid (WGRH) panels facilitate the efficient generation of high resolution genomic maps using both polymorphic and non-polymorphic markers. The hybrid cells are created by lethally irradiating the donor bovine cell line using X-rays to fragment the chromosomes (to
about 10Mb average size). These cells are then fused to a recipient hamster cell line, which is thymidine kinase deficient (TK-). Culturing the fused cells in media containing HAT (hypoxanthine, aminopterin, thymidine) ensures that only hybrid hamster cells containing bovine chromosomal fragments will grow - see schematic illustration (Gross and Harris 1975, Walter et al 1994).
The average retention of bovine DNA in the hybrid cells is tested by screening for presence of bovine genetic loci which have been mapped to different chromosomes. The hybrid cells are also tested using fluorescent in situ hybridisation (FISH). This enables the bovine content and average
DNA fragment size of randomly chosen hybrids to be estimated - see photo. The resolution of the hybrid panel depends on fragment size which is proportional to the initial radiation dose, and to a lesser extent, the retention frequency of the bovine DNA fragments in the hybrids.Using an initial
X-ray dose of 500 rads, a map resolution of at least 500Kb can be expected which is significantly higher than any bovine map currently available. As the bovine DNA is on a background of hamster chromosomes, non-polymorphic markers are as informative as genetic markers, making this an ideal resource for mapping ESTS.

By directly mapping a large number of bovine expressed sequences, and building comparative mapping links to identify comparative positional candidate genes, a high resolution whole genome radiation hybrid map in cattle will provide the tools to efficiently identify candidate genes for disease and quantitative traits.

WGRH panels - the choice for high resolution genome mapping

Schuler et al (1996) have placed over 16,000 human cDNAs on the human map using the Genebridge4 and G3 WGRH panels. Stewart et al (1997) mapped over 10,000 STSs to the human G3 panel. The G3 panel was generated using 10,000rads of X-rays and has an estimated potential
resolution of 240kb, with an average DNA fragment size of 3Mb. Although such a high radiation dose offers a high resolution map, many thousands of STSs are required to generate the framework map. Even with 10,478 markers, there are still gaps in the G3 map. There are not this many markers available for the bovine genome, so generating a WGRH panel of this resolution would be premature. In order to build a framework map using 1000-2000 markers, a lower initial radiation dose is necessary. The framework map for the 3000rad human Genebridge 4 panel was generated using 450 carefully chosen markers. We propose to generate a panel with a resolution intermediate between these two human panels. A 5000rad bovine WGRH panel should provide a map resolution of at least 500Kb which is equivalent to 0.5cM assuming a map length of about 3000cM.

In addition to building a high density map of the bovine genome itself, the WGRH panel can be used to generate a high resolution comparative map, by mapping genes present on maps from other species.

The WGRH panel also provides a means of identifying and integrating large fragment clones, such as YACs or BACs with linkage and physical maps and to assign ESTS to specific clones. Management of the RH panel generated through this project would be integrated with the bovine YAC and BAC libraries already constructed.

Contact Dr John Williams for more information.

Pictures adapted from "Radiation Hybrid Mapping in Pigs" poster presented at 13th European Colloquium on Cytogenetics of Domestic Animals.
Hungary Academy of Sciences, Budapest, Hungary, 1-6 June, 1998.
Néstor López-Corrales, Chris Mungall, Alan L. Archibald - Roslin Institute
Linda McCarthy, Sharon McDowell and Peter Goodfellow - University of Cambridge.