Genetic architecture of adiposity in the cross of LG/J and SM/J inbred mice

Abstract

The genetic basis of variation in obesity in human populations is thought to be owing to many genes of relatively small effect and their interactions. The LG/J by SM/J intercross of mouse inbred strains provides an excellent model system in which to investigate multigenic obesity. We previously mapped a large number of quantitative trait loci (QTLs) affecting adult body weight in this cross. We map body composition traits, adiposity, and skeletal size, in a replicate F2 intercross of the same two strains containing 510 individuals. Using interval-mapping methods, we located eight QTLs affecting adiposity (Adip1-8). Two of these adiposity loci also affected tail length (Adip4 and Adip6) along with seven additional tail length QTLs (Skl1-7). A further four QTLs (Wt1-4) affect adult weight but not body composition. These QTLs have relatively small effects, typically about 0.2-0.4 standard deviation units, and account for between 3% and 10% of the variance in individual characters. All QTLs participated in epistatic interactions with other QTLs. Most of these interactions were due to additive-by-additive epistasis, which can nullify the apparent effects of single loci in our population. Adip8 interacts with all the other adiposity QTLs and seems to play a central role in the genetic system affecting obesity in this cross. Only two adiposity QTLs, Adip4 and Adip6, also affect tail length, indicating largely separate genetic control of variation in adiposity and skeletal size. Body size and obesity QTLs in the same locations as those discovered here are commonly found in mapping experiments with other mouse strains.

Characterization of a QTL affecting skeletal size in mice

“…The estimated location of this QTL is proximal to that of the QTL described in this study, and the confidence intervals from the two studies only just overlap. The Chr 1 QTL described by Cheverud et al (2001) affects tail length in both sexes, as in this study, but does not have an effect on body mass at 10 weeks of age, consistent with the relatively weak or non-significant effects on body mass observed in this study. Klingenberg et al (2001) and performed genome-wide scans for QTL affecting mandible size and femur length, respectively (among other traits), but did not detect QTL for these traits on Chr 1.…”

“…In an intercross between two inbred strains selected for high and low body size, Cheverud et al (2001) identified seven QTL affecting tail length, one of which was located on Chr 1. The estimated location of this QTL is proximal to that of the QTL described in this study, and the confidence intervals from the two studies only just overlap.…”

Contribution of maternal effect QTL to genetic architecture of early growth in mice

“…The probability associated with the multiple regression at each QTL position was transformed to a linear scale by logarithmic transformation [LPR = log 10 (1/Prob.)] (see Cheverud et al, 2001) and, as a result, our LPR scores can be directly compared to LOD scores obtained through maximum likelihood methods (Lander and Botstein, 1989). Environmental covariates were included in the mapping to remove phenotypic variation.…”

“…The 5% chromosome-wide significance levels were obtained by dividing 0.05 by the effective number of markers on that chromosome. For a discussion of the merits of using a chromosome wide significance level see Cheverud et al (2001). The genome-wide significance level was obtained by dividing 0.05 by the effective number of markers in the genome.…”

“…These correlations suggest some physiological relationship between expression of some maternal characters and maternal effects. Because the individuals used in this analysis have also been used in mapping studies of body size, and adiposity (Vaughn et al, 1999;Cheverud et al 2001) we can examine whether the maternal effect QTL map to the same genomic locations as loci influencing these other characters. We find that one maternal effect QTL (D6Mit9+10) maps close to a locus influencing adiposity and shows the same pattern of a positive additive effect and negative dominance effect, suggesting a single locus that influences both characters through a similar mode.…”