Copy number variation in the mouse genome: implications for the mouse as a model organism for human disease

Abstract

Individuals within a species have genetic differences which ultimately result in the spectrum of phenotypic variation that we observe. Genetic variation exists at the nucleotide level in the form of single nucleotide polymorphisms (SNPs), and at a structural level as inversions, deletions and amplifications of larger stretches of nucleotides. Profiling of human and mouse genomes has identified numerous genomic segmental copy number variations (CNVs) throughout these genomes. Since inbred mice are widely used laboratory models for the study of both normal and disease biology, it is crucial that we understand the full scope of genetic variation, including CNVs, within these animals. These genetic differences can inform us about the history of a population or species, enlighten us on gene function, and guide our selection of a model system for the study of human disease.

Fig. 1.

A comparison of amplitude histograms of CNVs from Cutler and Graubert. The smoothed histograms of the log₂ fold-change values of the CNV datasets from Cutler (all significant CNVs; thick line) and Graubert (all CNVs with an absolute log₂ amplitude >0.5; dashed line) are shown. The positions of whole-number ratio changes in copy number are indicated in parenthesis on the X-axis. Deletions are truncated at a minimum fold-change of 1/20 (log₂ = −4.3). The two curves are scaled for plotting and the peak at −4.3 is off scale.

Fig. 2.

A ‘partial amplitude’ deletion on murine chromosome 14. The log₂ fold-change aCGH data for 42 inbred mouse strains compared to C57BL/6J is plotted for a region on chromosome 14 from nucleotide position 68206900 to 68418187. Data for C57BL/6J is shown in black while all other strains are shown in color. The locations of expected fold-changes based on whole-number copy losses are shown with dotted lines. The positions of known genes in this locus are shown at the top.

Fig. 3.

SNP- and CNV-based inbred mouse strain phylogenies. Trees generated by the SplitsTree4 hybridization tree-algorithm are shown. Major sections of the SNP-based tree are colored to highlight their locations. This same strain coloring is used for the CNV-based tree. Branch lengths within a tree are proportional to phylogenetic distance, and those with discontinuity markers are actually twice the length shown. Black lines represent regions of hybridization within the trees, where genetic information is being exchanged between lineages or their precursors.

Fig. 4.

The rate of CNV accumulation between related strain pairs. The ratio of the number of CNVs which vary between a given pair of strains and the number of years of separation of that pair are shown for 11 inbred mouse strain pairs. The mean across these values (0.37 CNVs/year) is indicated by a dashed line. The strain pairs used are: 1) 129S1/SvImJ and 129X1/SvJ (55 years separation); 2) C3H/HeJ and CBA/J (88 years); 3) C57BL/6J and C57BL/10J (71 years); 4) C57BR/cdJ and C57L/J (75 years); 5) C57L/J and C57BL/6J (87 years); 6) C57L/J and C57BL/10J (87 years); 7) C57BR/cdJ and C57BL/6J (87 years); 8) C57BR/cdJ and C57BL/10J (87 years); 9) DBA/1J and DBA/2J (79 years); 10) FVB/Ntac and SWR/J (82 years); 11) FVB/Ntac and SJL/J (82 years).

Fig. 5.

CNVs at the Klra locus. The log₂ values for the aCGH data (points) and the locations of called CNVs (boxes) are shown for C57BL/6J and all 20 strains with CNVs in the region of chromosome 6 from nucleotide 129787139 to 103362567. Amplification CNVs and aCGH values within them are shown in red and deletion CNVs and their data points are shown in green. Also indicated are the 3-point running means of the aCGH data as a black line for each strain. At the bottom, the exon (thick line) and intron (thin line) positions for all the Klra genes in this locus are shown in red.