The genome of the laboratory mouse is thought to be a mosaic of regions with distinct subspecific origins. We have developed a high-resolution map of the origin of the laboratory mouse by generating 25,400 phylogenetic trees in 100 kb intervals spanning the genome. On average 92% of the genome is of M. m. domesticus origin and the distribution of diversity is strikingly non random among the chromosomes. There are large regions of extremely low diversity, representing blind spots for studies of natural variation and complex traits, as well as hot spots of diversity. In contrast with the mosaic model we found that the majority of the genome has intermediate levels of variation of intrasubspecific origin. Finally, the wild-derived mouse strains that are supposed to represent different mouse subspecies show substantial intersubspecific introgression. This has serious implications for evolutionary studies that assume these are pure representatives of a given subspecies.Laboratory mice, the most popular model organism in mammalian genetics 1,2 , were derived from wild mice belonging to the Mus musculus species by an intricate process that included the generation of "fancy" mice in both Asia and Europe and a complex web of relationships among inbred strains 3 . Early studies demonstrated that the mitochondria and the Y chromosome present in many classical laboratory strains were derived from different subspecies, M. m. domesticus for the mitochondria and M. m. musculus for the Y chromosome 4,5 . Furthermore, the Y chromosome was introduced in the laboratory mouse through M. m. molossinus males 6 . Based on these findings, it was proposed that the genomes of inbred strains were a mosaic of regions with different subspecific origin 7 . Recently, the fine structure of such mosaic variation has been described 8 . This study reported that strain-to-strain comparisons revealed regions with extremely high variation spanning one third of the genome and regions with extremely low variation covering the remaining two thirds of the genome. This distinctively bimodal distribution was assumed to represent regions with the same and different subspecific origin. This mosaic model has been the driving concept behind mouse association mapping studies and haplotype analysis [9][10][11][12] . However, the origin of a given region of a laboratory strain could not be directly assigned to a subspecies due to the lack of reference sequences for the three major mouse subspecies. Subsequent studies raised questions regarding the haplotype structure 11,13 , the effect of ascertainment biases in subspecific assignment [14][15][16] and the contributions of intersubspecific