Experiment
To determine the number and strength of genetic variants that modify Zfp423nur12/nur12 cerebella phenotype a (129S1/SvImJ x C.B6-Zfp423nur12)F2-Zfp423nur12/Zfp423nur12 intercross was performed.
220 homozygotes were dissiected, photographed and assigned categorical phenotypes.
Phenotype 1 characterized by no evident cerebllum;
Phenotype 2 characterized by rudimentary flaps of cerebella tissue;
Phenotype 3 characterized by hemispheres that do not meet at the midline;
and Phenotype 4 characterized by hemispheres that meet at the midline.
Of the 220 mice sacrificed 45 had phenotype 1, 51 had phenotype 2, 56 had phenotype 3 and 70 had phenotype 4.
A whole genome linkage scan on the DNA from these mice was performed using SSLP markers with an average spacing of 14.8 cM. This analysis identified 3 significant linkage peaks on Chrs 2, 3 and 17 as well as suggestive linkage on Chr 11 and 15.
Next the categorical phenotypes were reevaluated as one of 3 dichotomous traits under a binary model.
Dichotomous trait 1 was defined as the absence of cerebellum versus any visible cerebellar tissue ( phenotype 1 versus 2, 3, and 4). For this trait, nominally significant loci were found on Chr 12 and at the Chr 15 locus suggested by the catergorical analysis.
Dichotomous trait 2 contrasted minimal cerebellar tissue with organized hemispheres (phenotype 1,2 versus 3,4). Significant linkage was found only at the categorical trait loci on Chr 2 and Chr 17.
Dichotomous trait 3 contrasted midline adjacent hemispheres with all others (phenotypes 1, 2, 3 versus 4). A single significant peak was detected at the Chr 3 locus.
Genome-wide significance of the linkage statistics survived correction for the number of phenotype models for the QTL detected on Chrs 2, 3, 15 and 17. These QTL were referred to as Anatomic modifiers of Zfp423nur12 - Amzn QTL.
The 4 significant Amzn QTL were further refined by typing finely spaced genetic markers through each interval at each linkage peak, considering phenotypic direction and dominance.
The Chr 3 QTL Amzn1, anatomical modifier of Zfp423 1, LOD=4.62, p=0.0019 mapped to 19.5 cM for the categorical trait and with LOD=6.5, p<10-5, at 22.0 cM for the dichotomous trait 3. Peak marker was D3Mit25, within 95% BCI for both traits. The allele effect showed reduced hemispheres in 129S1 genotypes. Heterozygote phenotypes were significantly different from both parental homozygotes suggesting substantial but incomplete dominance of the 129S1 allele at this locus.
The Chr 17 QTL Amzn2, anatomical modifier of Zfp423 2, was significant for both categorical, LOD=4.82, p=0.0013 and dichotomous trait 2, LOD=4.99, p=0.0012, at intervals 34.0 and 28.0 cM , respecitvely. Peak linkage was at D17Mit139 for both traits. The allele effect was towards reduced hemispheres in BALB homozygotes. The 129S1 allele appeared dominant with heterozygote phenotypes significantly different from BALB but not from 129S1.
The Chr 15 QTL Amzn3, anatomical modifier of Zfp423 3, was significant for only dichotomous trait 1, LOD=4.63, p=0.0054; peaking at marker D15Mit270 between 31.6 and 12.0 cM. The allele effect was towards reduced hemispheres in 129S1. Heterozygotes were significantly different from 129S1 but only nominally different from BALB suggesting a BALB dominant model.
The Chr 2 QTL, Amzn4, anatomical modifier of Zfp423 4, with LOD=4.01, p=0.013 for both categorical trait (38.0 cM) and dichotomous trait 2 (40.0 cM) peaked with marker D2Mit151.
The allele effect was toward reduced hemispheres in 129S1 genotypes. Heterozygous phenotypes were significantly different from 129S1 but not from BALB, consistent with a BALB dominant model at this locus.
Seeking to detect synthetic interactions between the QTL, to better explain the proportion of phenotypic variance, the one-dimensional linkage analysis was repeated using each of the previously identified QTL as a covariate. The results showed significant interactions with the identified Amzn QTL, these were referred to as synthetic modifiers of Zfp423nur12.
For the categorical trait significant interaction was seen with Amzn1 as the covariate on Chr 2, peaking at 69.7 cM, with a 95% BCI mapping between 57.7-105.7 cM, LOD=3.46, p=0.0163, accounting for 10.8% of variance. The locus was labeled Smzn1a, synthetically interacting modifier of Zfp423 1A.
For the categorical trait significant interaction was also seen with Amzn2 as the covariate on Chr 13, peaking at 23.1 cM, with a 95% BCI mapping between 17.3-49.3 cM, LOD=3.42, p=0.0332, accounting for 12.3% of variance. This locus was labeled Smzn2a, synthetically interacting modifier of Zfp423 2A.
For dichotomous trait 1 Amzn3 showed significant interaction with Chr 2, peaking at 105.7 cM, with a 95% BCI mapping between 81.7-105.7 cM, LOD=3.40, p=0.0439, accounting for 14.2 % of the variance. This locus was labeled Smzn3, synthetically interacting modifier of Zfp423 3.
For dichotomous trait 2 Amzn2 showed significant interaction with three Chromosomes:
Smzn2b, synthetically interacting modifier of Zfp423 2B, on Chr 2, peaking at 51.7 cM, with a 95% BCI mapping between 35.7-63.7 cM, LOD=3.50, p=0.0292.
Smzn2c, synthetically interacting modifier of Zfp423 2C, on Chr 13, peaking at 25.3 cM, with a 95% BCI mapping between 17.3-45.3 cM, LOD=3.25, p=0.0475.
Smzn2d, synthetically interacting modifier of Zfp423 2D, on Chr 14, peaking at 67.3 cM, with a 95% BCi mapping between 55.3-70.6 cM, LOD=3.30, p=0.0442; all three loci accounting for 21.0% of trait variance.
For dichotomous trait 3 Amzn1 interacted as a covariate with Chr 7, peaking at 76.9 cM, with a 95% BCI mapping between 69.5-76.9 cM, LOD=3.40, p=0.0327, accounting for 18.7% of variance. This locus was labeled Smzn1b, synthetically interacting modifier of Zfp423 1B.
Analysis Tools