In the final example, we tested the effect of CNVs on gene expression and phenotypes. A CNV region on chromosome 4: 62.49–62.52 Mb that spans both Alad and Hdhd3—is interesting and involves a 4 × expansion in strains with the D haplotype. The 30 kb CNV is otherwise identical by descent (Fig. 6a). This CNV is linked with high variation in mRNA expression of Alad and Hdhd3 in multiple brain regions (Fig. 6b,c), lung (q=2.1 × 10−7), eye (q=1.3 × 10−10) and liver (q=9.2 × 10−4). Quantitative proteomics of hippocampus confirms significant upregulation (ALAD 2.3-fold, P<0.01, HDHD3 1.5-fold, P<0.01, see Supplementary Data 13). The CNV expansion of Alad and Hdhd3 is strongly linked to two classic phenotypes: pain response (GN ID 11307; q=7.8 × 10−3) and deoxycorticosterone levels in cerebral cortex (GN ID 12568; q=2.6 × 10−4) (Fig. 6d). A matched phenome scan in human demonstrates that rs1800435 in ALAD is associated with chronic pain (P=2.2 × 10−2; logistic regression) (Fig. 6e).
