In silico QTL mapping of maternal nurturing ability with the mouse diversity panel

Abstract

Significant variation exists for maternal nurturing ability in inbred mice. Although classical mapping approaches have identified quantitative trait loci (QTL) that may account for this variation, the underlying genes are unknown. In this study, lactation performance data among the mouse diversity panel were used to map genomic regions associated with this variation. Females from each of 32 inbred strains (n = 8-19 dams/strain) were studied during the first 8 days of lactation by allowing them to raise weight- and size-normalized cross-foster litters (10 pups/litter). Average daily weight gain (ADG) of litters served as the primary indicator of milk production. The number of pups successfully reared to 8 days (PNUM8) also served as a related indicator of maternal performance. Initial haplotype association analysis using a Bonferroni-corrected, genome-wide threshold revealed 10 and 15 associations encompassing 11 and 13 genes for ADG and PNUM8, respectively. The most significant of these associated haplotype blocks were found on MMU 8, 11, and 19 and contained the genes Nr3c2, Egfr, Sec61g, and Gnaq. Lastly, two haplotype blocks on MMU9 were detected in association with PNUM8. These overlapped with the previously described maternal performance QTL, Neogq1. These results suggest that the application of in silico QTL mapping is a useful tool in discovering the presence of novel candidate genes involved in determining lactation capacity in mice.

Fig. 1.

Average daily litter gain (ADG) and pup survival among 32 inbred mouse strains. Females from each strain were studied in their 2nd lactation. Females were induced to ovulation by treatment with PMS and HCG and then mated to CD-1 males. On day 1 postpartum, number of pups born per litter was recorded and independent cross-foster litters of 10 1-day-old CD-1 pups were then given to each dam. ADG weight gain (A) and maternal litter rearing capacity (B) were recorded daily through day 8 postpartum (PNUM8). Each symbol represents the mean ± 95% confidence interval (CI) for 8–16 dams with the exception of CAST/EiJ for which only a single dam was obtained. Numbers in parentheses indicate the number of animals studied. The solid gray lines indicate the 95% CIs about the overall mean for each trait. Symbols are color coded by the percentile ranking of the strain mean for ADG.

Fig. 2.

Comparison of genetic relatedness with ADG and maternal pup survival among 32 inbred mouse strains. A dendrogram (left) illustrating the genetic relatedness among the strains used in this study is presented alongside a heat map (middle) illustrating the relative ranking for ADG and PNUM8. An additional heat map (right) illustrates the pairwise relatedness among the strains expressed as a Pearson's distance (d = 1 − r, where r is Pearson's correlation). The dendrogram is a UPGMA tree based on average haplotype block identity across the entire genome. Pair-wise genetic distances between strains were computed as described under genetic similarity in materials and methods.

Fig. 3.

Manhattan plots illustrating significant genome-wide associations for ADG and PNUM8. The graphs present the results of haplotype association analysis for ADG (A) and PNUM8 (B). Each bar represents the location in the genome that begins a haplotype block. The significance threshold for a Bonferroni-corrected genome-wide (GW05) P value of α = 0.05 (2.75 × 10⁻⁶) is indicated by the dashed line.

Fig. 4.

Definition of haplotype blocks for MMU11 that are in close proximity to the genes for Sec61g and Egfr. Single nucleotide polymorphism (SNP) data for MMU11 were loaded into Haploview to confirm the block structure in this region and define the major haplotypes for the region. These same data were also used in an independent single SNP analysis using the efficient mix-model analysis server (http://mouse.cs.ucla.edu/emmaserver/). Association score is plotted against the position for the start of each of 977 haplotype blocks that were defined on this chromosome (A). The region from 15 to 17 Mbp is enlarged to illustrate highly significant single SNP associations in relationship to known genes and linkage disequilibrium (LD), which is expressed as r² in the heat map (B).

Fig. 5.

Definition of haplotype blocks on MMU19 that are in close proximity of the gene for Gnag. SNP data for MMU19 were loaded into Haploview to confirm the block structure in this region and define the major haplotypes for the region. This same data were also used in an independent single SNP analysis using the efficient mix-model analysis server (http://mouse.cs.ucla.edu/emmaserver/). Association score is plotted against the position for the start of each of 473 haplotype blocks that were defined on this chromosome (A). The region from 15 to 17 Mbp is enlarged to illustrate highly significant single SNP associations in relationship to known genes and LD, which is expressed as r² in the heat map (B).

Fig. 6.

Definition of haplotype blocks on MMU8 that are in close proximity to the gene for Nr3c2. SNP data For MMU8 were loaded into Haploview to confirm the block structure in this region and define the major haplotypes for the region. This same data were also used in an independent single SNP analysis using the efficient mix-model analysis server (http://mouse.cs.ucla.edu/emmaserver/). Association score is plotted against the position for the start of each of 976 haplotype blocks that were defined on this chromosome (A). The region from 15 to 17 Mbp is enlarged to illustrate highly significant single SNP associations in relationship to known genes and LD, which is expressed as r² in the heat map (B).

Fig. 7.

Analysis of mammary mRNA abundance for Egfr, Nr3c2, and Gnaq indicates a low negative correlation with ADG during an established lactation. Total RNA was prepared from mammary tissue collected at the end of the study period on day 10 postpartum from each of 32 inbred strains. Real-time RT-PCR was then conducted with primers specific for Egfr (A), Nr3c2 (B), Gnaq (C), and Sec61g (D). Nucleolin and Ppia mRNAs were measured as reference transcripts. Data are expressed at △ct. Each symbol represents the mean ± 95% CI for 5 animals (with the exception of CAST/EiJ, which had only 1). Symbols are color coded by ADG percentile ranking (see Fig. 1).