Here we present a strategy to determine the genetic basis of variance in complex phenotypes that arise from natural, as opposed to induced, genetic variation in mice. We show that a commercially available strain of outbred mice, MF1, can be treated as an ultrafine mosaic of standard inbred strains and accordingly used to dissect a known quantitative trait locus influencing anxiety. We also show that this locus can be subdivided into three regions, one of which contains Rgs2, which encodes a regulator of G protein signaling. We then use quantitative complementation to show that Rgs2 is a quantitative trait gene. This combined genetic and functional approach should be applicable to the analysis of any quantitative trait.

We describe three patients from two families with progressive spinocerebellar ataxia, peripheral neuropathy, raised alpha-fetoprotein (AFP) and hypercholesterolaemia. Two siblings had identical clinical features, with late childhood onset of symptoms and slow progression, requiring crutches to walk at ages 37 and 38 years. Another patient developed ataxia aged 13 years and became wheel-chair bound by 20 years of age. Although they all had raised serum AFP levels, their clinical, immunological, biochemical, cytogenetic and molecular genetic studies failed to support a diagnosis of Ataxia Telangiectasia. Extensive investigation including imaging, biochemical and genetic studies excluded other known ataxias. Their clinical features most closely resemble the phenotype of a single consanguineous Japanese family with four individuals affected by spinocerebellar ataxia, peripheral neuropathy, raised AFP and hypercholesterolaemia. Homozygosity mapping has identified a locus in this Japanese family at 9q34. Haplotype analysis of our cases demonstrated possible linkage to 9q34, suggesting these may be the first Caucasian families described with this disorder.