Identify a set of single nucleotide polymorphisms (SNPs) that span the mouse genome at high resolution (ca. 1 per 50-100,000 bp), building a platform upon which to implement efficient, array-based genome scanning methods for identifying genes that influence cancer phenotypes.
A. Rationale:
The lessons of human cancer genetics tell us that genes altered in cancer cells are of crucial importance for understanding the molecular basis of cancer development and progression. Further, there is ample reason to suspect that modifier genes conveying resistance or susceptibility toward specific cancers will prove crucial for understanding the etiology of human cancers, and that other modifier genes may affect responses to particular prevention and therapeutic strategies, and hence that patient genotype could guide clinical practice. The convenient genetic manipulatability of the mouse, and an increasing body of evidence that mouse cancers show both genomic alterations and responsiveness to modifier genes, makes a strong case that mouse models of cancer represent fertile ground for identifying new genes altered in particular cancers, and perhaps more importantly, avenues to map, identify, and study the functional contributions of modifier genes which are not themselves altered, but which profoundly affect cancer phenotypes. There are two big impediments to efficiently identifying both classes of genes (altered in cancers, and modifiers) in the ensemble of mouse models of human cancer which are or will become available in the next few years: 1) identifying genetic loci by LOH or linkage analysis and then mapping them at high resolution onto the mouse genome; and 2) positionally cloning the localized regions and identifying the involved gene. This recommendation is focused on alleviating the first impediment, namely the mapping of cancer gene loci, which currently involves cumbersome techniques based on simple sequence repeat polymorphisms (SSRP). While tractable by committed mouse genetics labs, SSRP-based genomic scanning and high resolution mapping is too expensive and time consuming for the typical cancer biology lab; indeed, only a handful of the hundreds of current mouse cancer models are being assessed for affected genes or modifier genes.
The consensus of this committee is that a technology exists to dramatically change that perspective, allowing efficient genome scanning and thus the realistic possibility that virtually every meritorious model of cancer will be assessed for both classes of cancer genes. The enabling technology involves the use of single nucleotide polymorphisms (SNPs) instead of SSRPs. In conjunction with some form of array based platform for parallel processing to display SNPs, it should be possible to scan a cell's genome in one or a few assays, instead of hundreds or thousands of SSRP analyses. In turn, the knowledge of synteny and of the increasingly high resolution EST maps for human and mouse will likely produce candidate genes that in many cases should obviate positional cloning procedures. There has already been preliminary 'proof-of-principle' for the applicability of human SNPs for mapping, in particular in the context of microchip arrays. Moreover, the pace of discovery and development in the area of array technologies argues that there will be options of increasing power and decreasing cost, all based on SNPs as a discriminator. Thus, the key element is a dense set of SNPs mapped onto the mouse genome, not the specific choice of a particular display technology, since the SNPs will be the generic component. Therefore, we propose below that a SNP Map of the mouse genome be developed as quickly as possible, placed in the public domain, and then made available on non-exclusive basis to be incorporated into the development of array technologies that would render genomic scanning and high resolution mapping tools readily available to the research community studying mouse models of cancer.
top
B. Specific Recommendations:
B.1. Encourage the development and refinement of technologies (a 'SNP Detector') for efficiently identifying mouse SNPs across a set of mouse strains and species used in cancer research. There are probably 8-10 'core' mouse strains that show differential cancer phenotypes or are otherwise commonly used in building animal models of cancer. In addition, dozens of other strains and species may harbor modifier genes, presumably syntenic to ones segregating in the human population, that provide resistance or susceptibility (or otherwise alter) to particular cancer pathways. Thus, there is clear merit in developing a 'SNP-Detector' that can eventually score not only a core set of strains, but any and all others for which there are clues to warrant their use in genetic mapping studies. The NCI/NHGRI/NIH should therefore encourage technology development of SNP-Detectors that can over time assess/extend the 'Mouse SNP Library' to incorporate all strains of interest. There have been technological demonstrations of such scanning-for-SNP capabilities, and as such reason to believe that a SNP-Detector can be developed to realize this goal. (Notably, this goal may be interfacable with an ongoing RFA that seeks to expedite the discovery of human SNPs.)
B.2 Elicit the generation and dissemination of a Mouse SNP Map that spans the mouse genome, with SNPs spaced every 50-100 kb, or about two per BAC clone of the projected physical map of the mouse genome (for the 'core strains'). The opportunity presented by high resolution SNP Maps has already been set for human, and the path forward is clear for this model organism. There should be a call to produce the mouse SNP Map, initially for a core set of inbred strains, perhaps comparing different SNP-Detector technologies. Then, as a particularly successful SNP-Detector technology is identified, the machine should be ramped up to produce the high resolution Mouse SNP Map, both for the core strains, and incorporating other meritorious strains and species. The principle should be that the SNPs are immediately posted on the WEB and into the public domain, and that regardless of the SNP-detector technology used, the Mouse SNP Map should be available to all potential users, both of the SNPs themselves, and for incorporation into various formats for highly parallel screening.
B.3 In various partnerships, seek to have the Mouse SNP Map incorporated into arrays and other formats that would allow high throughput genome scanning and fine structure linkage analysis so as to motivate widespread use of genetic mapping technologies, consequently expediting the discovery of genes which influence cancer phenotypes in mouse models of human cancers. Our view is that the incorporation of SNPs into arrays or other highly parallel processes is independent of the Mouse SNP Map itself, and further that the area is undergoing intense technology development and refinement. Thus, we see no reason to chose a particular array technology, but rather propose the NIH encourage alternatives and their comparison by freely providing the Mouse SNP Map as a platform, and furthermore working with the technology developers to encourage them to produce mouse SNP screening tools and disseminate them to the cancer research community. By comparing alternative technologies with the same SNP platform, it should be possible to determine which is most resilient and cost effective, particularly for the generic RO1-funded NIH investigator studying mouse models so as to enable their research.
top
