Maximising the potential of wild-derived laboratory mouse strains for medical research

Our key aim is to explore the relationship between genetic and medically relevant human disease phenotypes. One way to do this is to assess the genetic differences between long-established laboratory mouse strains. Wild-derived mouse strains display many important disease phenotypes such as resistance to various forms of cancer (e.g. liver, lung, and skin cancer), bacterial and viral infection. However, the foundation for studying the genetic differences in these strains is having accurate genome sequences. In this project, we will first generate genome sequences for the most commonly used wild-derived mouse strains and then use these sequences and knowledge of the gene structures to determine the genetic cause of observed phenotypic differences between these strains. By combining sequence and phenotypic data we will determine whether sequence variants are likely to be contributing to disease susceptibility.

The 1.5-2M years of genetic divergence between M.m domesticus and the wild derived strains (SPRET/EiJ, PWK/PhJ, CAST/EiJ, and WSB/EiJ) makes it difficult to use standard methods of mapping back to the C57B6/J reference. We have observed 4-8 times more SNPs, short indels and larger structural variants (>100bp) in the wild derived strains so it is not appropriate to use the C57B6/J reference genome as a comparator for these strains. If we consider just protein coding regions of the genome alone, we can see that in Mus spretus just under 3,000 large structural variants occur in these regions and so are likely to result in significantly altered gene models. This problem poses a significant difficulty for any outbred crossing based projects (such the CC and the Diversity Outbred Cross (DO)) where a subset of or interspecific crosses where the founder strains are wild-derived. Without near contiguous sequences assemblies (at least in protein coding regions), we currently cannot accurately predict gene modes and hence predict function of polymorphism.

The task of generating accurate whole-genome draft assemblies of eukaryotic genomes from second-generation sequencing data remains challenging. The key ingredients required to generate high quality draft genome sequences of large eukaryotes are high sequencing depth from short fragment (200-600bp) paired-end sequencing libraries, lower depth from a range of large mate-pair fragment libraries (3-40Kbp), and a optical or physical map. This strategy has been used recently to produce several draft genome sequences for the goat, hamster, and gorilla.

The most obvious beneficiary of these genome sequences and annotation generated will be the mouse genetics community involved in mapping complex disease related traits, researchers mapping mutations in crosses involving the wild-derived strains and crosses attempting to identify modifiers of mutations. The wild-derived mouse strains are founder strains for two of the largest mouse recombinant inbred strain experiments, the Collaborative Cross and the Diversity Outbred Cross, with the numbers of DO mice that have been distributed by the Jackson Laboratory numbering into the thousands now. These lines have already been successfully used to investigate disease resistance and identify novel QTLs for fungal and viral infection response with many more disease resistance experiments expected to be carried out in the coming years.

The groups described above will benefit from this research by having a completely open access to the genome sequences of the wild-derived mouse strains and perhaps most importantly the strain specific gene annotation. This will mean that groups involved in mapping disease phenotypes in wild-derived mouse strains will be able to identify the host resistance genes involved in the phenotype as the first step to elucidating pathway information.