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
The microbiome influences health and disease through complex networks of host genetics, genomics, microbes and environment. To examine the interplay among the whole host genome, transcriptome and microbiome, the authors mapped QTL and correlated the abundance of cecal mRNA, luminal microflora, physiology and behavior in a highly diverse Collaborative Cross (CC) breeding population. One such relationship, regulated by a variant on Chr 7, was the association of Odoribacter (Bactoroidales Porphyromonadeceae) abundance and sleep phenotypes. In a test of this association in the BKS.Cg-Dock7m+/+ Leprdb/J mouse model of obesity and diabetes, known to have abnormal sleep and colonization by Odoribacter, treatment with antibiotics altered sleep in a genotype dependent fashion.
CC mice were bred and housed at Oak Ridge National Laboratory (ORNL). Phenotyping and genotyping were performed in at least one breeding pair per line from generations G2:F5G2:F8 for QTL analysis. Most genotyped mice came from the G2:F5 generation, which are estimated to be 75% inbred. Of the 650 CC lines initiated, 414 lines with at least a single male or female survived to G2:F5 generation.
When grand-progeny were born, members of the CC breeding population were subject to a high-throughput analysis broadly reflecting behavior, morphology and physiology (Philip et al, 2011). Phenotypes included wildness, activity monitoring in open field, light/dark box, piezo sleep, hot plate, tail-clip, blood chemistry and cell counts, fasting glucose. At dissection metrics of body weight and length, organ and fat pad weights, and bone composition. All CC phenotyping data and protocols are publicly available at the Mouse Phenome Database (phenome.jax.org; accession Chesler3).
Adult mice were euthanized by carbon dioxide and the cecum dissected. The number of animals varied such that for microbial abundance, 108 male and 108 female unique strains were used, and for genotyping 102 male and 99 female mice were used. A custom array using the Illumina iSelect platform for the Infinium system was developed for SNP genotyping as previously reported (PHILIP et al. 2011). Briefly, this array was based upon a subset of the 11,969 SNPs from the NIEHS-Perlegen SNP combined panel(Yang et al. 2007). The set was designed to discriminate all eight founder haplotypes and was optimized so that for any SNP on the array the maximum density of informative markers was used.
Mapping of QTL was performed using a modified version of the DOQTL R package (Gatti D.M. 2014). An alternate model that used the CC mouse haplotypes and existing sequence data from the Sanger Mouse Genomes Project and other sources to infer individual mouse genotypes at all loci was used (Gatti D.M. 2014). This enables the application of Genome Wide Association (GWA) to precisely identify those SNPs that predict phenotypic variation.
Using previously reported phenotypes in the CC breeding population (Philip et al. 2011), the authors sought to correlate these with the microbial abundance of mice of the same line. There were a total of 205 animals with 123 phenotypic measures and 13,618 OTUs (operational taxonomic units, used to classify cecal mRNA). The authors computed the Kendall rank correlation coefficient (Kendalls Tau coefficient) for the behavioral phenotypes with OTUs using the cor.test() function in the R statistical framework.
The authors performed QTL mapping to identify host genetic loci accounting for heritable variation in microbial abundance. There were 18 statistically significant (q < 0.05) microbial abundance (Micab) QTL among the mapped microbial OTU abundances (MGSCv37; mm9):
Micab1 (microbial abundance of Clostridiales Ruminococcaceae Oscillibacter 1) maps from rs6275656 to rs3165368 on Chr 1 with a peak LOD score of 8.20 at rs32084678 (13,279,810 bp).
Micab2 (microbial abundance of Bacteroidales Porphyromonadaceae Paludibacter 2) maps from rs31431100 to rs37044521 on Chr 3 with a peak LOD score of 8.23 at rs31103355 (108,854,325 bp).
Micab3 (microbial abundance of Clostridiales Lachnospiraceae Marvinbryantia 3) maps from rs30552223 to rs30158956 on Chr 3 with a peak LOD score of 8.13 at rs30089246 (37,569,141 bp).
Micab4 (microbial abundance of Clostridiales Lachnospiraceae Roseburia 4) maps from rs27619452 to rs3685172 on Chr 4 with a peak LOD score of 9.57 at rs32690134 (136,028,098 bp).
Micab5 (microbial abundance of Coriobacteriales Coriobacteriaceae Enterorhabdus 5) maps from rs29633871 to rs6354701 on Chr 5 with a peak LOD score of 8.84 at rs6377391 (119,128,609 bp).
Micab6 (microbial abundance of Clostridiales Lachnospiraceae Sporobacterium 6) maps from rs32246505 to rs32318125 on Chr 5 with a peak LOD score of 8.42 at rs8265964 (138,359,981 bp).
Micab7 (microbial abundance of Bacteroidales Porphyromonadaceae Odoribacter 7) maps from rs33107817 to rs6373775 on Chr 7 with a peak LOD score of 8.84 at rs31494696 (77,651,351 bp). The NZO/HlLtJ allele is associated with increased abundance at Micab7.
Micab8 (microbial abundance of Clostridiales Ruminococcaceae Lactonifactor 8) maps from rs3661776 to rs6176297 on Chr 7 with a peak LOD score of 8.88 at rs47611520 (47,761,932 bp).
Micab9 (microbial abundance of Bacteroidales Porphyromonadaceae Odoribacter 9) maps from rs33107817 to rs6373775 on Chr 7 with a peak LOD score of 8.60 at rs31494696 (77,651,351 bp).
Micab10 (microbial abundance of Clostridiales Lachnospiraceae Anaerostipes 10) maps from rs6281843 to rs31252778 on Chr 8 with a peak LOD score of 9.61 at rs32936112 (47,123,375 bp).
Micab11 (microbial abundance of Clostridiales Incertae Sedis XIV Blautia 11) maps from rs6399870 to rs50110045 on Chr 8 with a peak LOD score of 9.49 at rs33429737 (31,919,239 bp).
Micab12 (microbial abundance of Clostridiales Clostridiaceae Caminicella 12) maps from rs30432532 to rs33695839 on Chr 9 with a peak LOD score of 8.29 at rs30372085 (80,440,479 bp).
Micab13 (microbial abundance of Erysipelotrichales Erysipelotrichaceae Turicibacter 13) maps from rs6338556 to rs6265280 on Chr 10 with a peak LOD score of 8.87 at rs29327022 (88,018,183 bp).
Micab14 (microbial abundance of Bacteroidales Bacteroidaceae Bacteroides 14) maps from rs6314621 to rs26972849 on Chr 11 with a peak LOD score of 8.14 at rs26971743 (58,783,410 bp).
Micab15 (microbial abundance of Clostridiales Lachnospiraceae Syntrophococcus 15) maps from rs31931586 to rs49819430 on Chr 15 with a peak LOD score of 9.47 at rs6388530 (93,634,974 bp).
Micab16 (microbial abundance of Bacteroidales Porphyromonadaceae Tannerella 16) maps from rs36280504 to rs30760881 on Chr 19 with a peak LOD score of 8.10 atrs30320578 (47,732,625 bp).
Micab17 (microbial abundance of Clostridiales Ruminococcaceae Hydrogenoanaerobacterium 17) maps from rs29276152 to rs29306363 on Chr X with a peak LOD score of 8.50 at rs6292190 (155,749,346 bp).
Micab18 (microbial abundance of Clostridiales Lachnospiraceae Lachnobacterium 18) maps from rs8255374 to rs31682358 on Chr X with a peak LOD score of 8.20 at rs6213950 (163,790,061 bp).
Analysis Tools