Reference
The Collaborative Cross (CC) is a large (~1,000 line) panel of recombinant inbred (RI) mouse strains being developed through a community effort (Churchill et al. 2004). The CC combines the genomes of eight genetically diverse founder strains - A/J, C57BL/6J, 129S1/SvImJ, NOD/ShiLtJ, NZO/HlLtJ, CAST/EiJ, PWK/PhJ, and WSB/EiJ - to capture nearly 90% of the known variation present in laboratory mice. CC strains are derived using a unique funnel breeding scheme. Once inbred, the RI CC lines can be used to generate thousands of potential 'outbred' but completely reproducible genomes through the generation of recombinant inbred crosses (RIX). The designation 'PreCC' is used to describe a mapping population of CC mice that is still at incipient stages of inbreeding.
CTC (2004), Churchill, G. A., et al.. The Collaborative Cross, a community resource for the genetic analysis of complex traits. Nat Genet. 36, 1133-7.
Experiment
Allergic asthma is a common airway disease that is characterized in part by enhanced airway constriction in response to nonspecific stimuli.
To identify QTL that regulate responses to a common human allergen, the authors applied a house dust mite mouse (HDM) model of allergic airway disease (AAD) to 146 incipient lines of the Collaborative Cross (CC) and the CC founder strains. They employed a longitudinal study design in which mice were phenotyped for response to the bronchoconstrictor methacholine both before and after HDM sensitization and challenge using whole body plethysmography (WBP).
The authors obtained 153 male PreCC#/Unc mice (ages 10-14 wks) from Oak Ridge National Laboratory (Chesler et al. 2008; Kelada et al. 2012, 2014; Rutledge et al. 2014). Each mouse was from an independent CC line that had undergone 5-14 generations of inbreeding. For each of the eight CC founder strains, the authors obtained 5-9 mice per strain.
Since each PreCC#/Unc mouse was unique (i.e., not yet fully inbred), the authors chose to evaluate effect of allergic inflammation on airway responsiveness to methacholine using a longitudinal study design in which WBP measurements were made before and after the induction of AAD by sensitization and challenge with HDM allergen.
Analysis of WBP data was focused on the derived parameter "enhanced pause" (Penh). The two baseline measurements of Penh were found to be reproducible on a mouse-by-mouse basis (as shown in Figure S1). As such, the two sets of baseline WBP data were averaged together to create a mean value of Penh per dose of methacholine. For mice for which only one baseline measurement was available, the single measurement was used as the mean.
A quadratic model was used to calculate the dose of methacholine required to produce a 50% increase in baseline Penh, which the authors refer to as the provocative concentration 150 (PC[150]).
PC[150] values for each mouse were calculated before and after HDM treatment, referred to as baseline PC[150] and final PC[150], respectively, and the change in PC[150] due to allergen (delta PC[150]) was calculated as baseline PC[150] final PC[150].
There was significant variation in methacholine responsiveness due to both strain and HDM treatment, as reflected by changes in Penh.
Genotypes for all of the PreCC#/Unc mice used in these experiments have been previously reported (Rutledge et al. 2014). Each mouse was genotyped at the University of North Carolina - Chapel Hill, using one of two Affymetrix SNP arrays (A or B) that were produced in development of the Mouse Diversity array (MDA) (Yang et al. 2009). After removing uninformative and poorly performing SNPs, these arrays contained 181,752 (A-array) and 180,976 (B-array) SNP assays, and the set of SNPs on each array did not overlap. Most mice (83%) were genotyped on the B-array and the remaining were genotyped on the A-array. These training arrays were annotated to NCBI Build 36 of the mouse genome, but the authors mapped QTL boundaries to Build 37 positions to integrate with other resources. NCBI Build 37 positions are reported in the results.
The most probable ancestor for each SNP in each mouse was estimated using the GAIN algorithm (Liu et al. 2010), and founder haplotypes were reconstructed based on these results. The nonoverlapping SNP datasets from arrays A and B were merged by imputing unobserved genotypes based on inferred founder haplotype, resulting in ~360,000 SNP markers in total. The authors used HAPPY (Mott et al. 2000) to infer ancestry matrices for an additive genetic model for QTL mapping. For efficiency, the matrices were averaged across SNPs between which GAIN inferred no recombination in the population, and this reduced the mapping dataset to 27,059 intervals with an average spacing of 95 kilobases. BAGPIPE (Valdar et al. 2006) was used to fit a regression model for log-transformed PC[150] and report LOD scores. Significance thresholds were determined by permutation (n = 10,000). The 1.5 LOD drop method was used to approximate confidence intervals for QTL (Dupuis and Siegmund 1999). The percent of phenotypic variation explained by each QTL was estimated by regression of phenotypes on haplotype probabilities at the peak locus.
QTL mapping was performed using log (ln)-transformed values of baseline PC[150] and final PC[150], and untransformed values of delta PC[150]. Four QTL were identified:
QTL Mcsq1 (methacholine sensitivity QTL 1, baseline) was significant for baseline measurements of PC[150]. It mapped to Chr18:63.670-65.295 Mbp with a peak LOD score of 6.95 at 64.761 Mbp and accounted for 20% of the phenotypic variance.
QTL Mcsq2 (methacholine sensitivity QTL 2, baseline) was suggestive for baseline measurements of PC[150]. It mapped to Chr14:21.810-28.728 Mbp with a peak LOD score of 6.14 at 22.960 Mbp.
QTL Mcsq3 (methacholine sensitivity QTL 3, final) was significant for final measurements of PC[150]. It mapped to Chr19:27.566-31.134 Mbp with a peak LOD score of 7.62 at 29.639 Mbp and accounted for 21% of the phenotypic variance.
QTL Mcsq4 (methacholine sensitivity QTL 4, final) was suggestive for final measurements of PC[150]. It mapped to Chr13:18.322-39.773 Mbp with a peak LOD score of 6.14 at 23.826 Mbp.
No QTL were detected for delta PC[150], and no alleles were conclusively identified for any of the mapped QTL (although A/J, C57BL/6J, and 129S1/SvImJ were discussed).
Analysis Tools