RT Journal Article
SR Electronic
T1 Genome-wide association study, replication, and mega-analysis using a dense marker panel in a multi-generational mouse advanced intercross line
JF bioRxiv
FD Cold Spring Harbor Laboratory
SP 387613
DO 10.1101/387613
A1 Xinzhu Zhou
A1 Celine L. St. Pierre
A1 Natalia M. Gonzales
A1 Riyan Cheng
A1 Apurva Chitre
A1 Greta Sokoloff
A1 Abraham A. Palmer
YR 2018
UL http://biorxiv.org/content/early/2018/08/08/387613.abstract
AB Genome-wide association studies (GWAS) in multigenerational outbred populations offer improved mapping precision compared to traditional populations such as F2s. Advanced intercross lines (AILs) are the simplest possible multigenerational intercross; AILs are produced by crossing two inbred strains and then breeding unrelated offspring for additional generations. New recombinations accumulate with each successive generation; these new recombinations provide increased mapping resolution. We used genotyping-by-sequencing (GBS) to re-genotype a cohort of sparsely genotyped F34 LG/J x SM/J AIL mice that were the subject of several prior publications as well as to obtain genotypes for new cohort of AIL mice from the F39-43 generations. The denser set of GBS markers allowed us to identify 110 significant loci, 36 of which were novel, for 79 behavioral and physiological traits in the 428 F34 mice. Genetic correlations between F34 and F39-43 were high, though F39-43 AILs showed systematically lower SNP-heritability estimates. We explored replication of loci identified in either F34 or F39-43 in the other cohort for the traits measured in both F34 and F39-43: locomotor activity, body weight, and coat color. While coat color loci were robustly replicated, we observed only partial replication of associations for locomotor activity and body weight. We then performed a mega-analysis of locomotor activity, body weight by combining F34 and F39-43 mice (N=1028), which identified four novel loci. Finally, we showed that imputation using the old, sparse marker set and the newer dense marker set had little impact on our results, emphasizing the need of the denser GBS genotypes. The present study provides empirical insights into replication, the utility of denser genotyping and identifies new candidate loci that can be explored in future studies.Author summary Using a sufficiently dense marker set is essential for genome-wide association studies. The required SNP density is a function of population history. When considering populations derived from a cross between two or more inbred strains, more recombinations will require a correspondingly greater number of markers. We performed several genome-wide association studies using different marker sets - sparse SNPs from a previously publish dataset, dense SNPs obtained using genotyping-by-sequencing, and imputed versions of both the sparse and dense SNPs - to compare their association results in 34th and 39-43rd generations of an LG/J x SM/J advanced intercross line (AIL). We found that the dense SNP set substantially increased the number of significant results, while the imputed sets did not. To achieve maximum power, we performed a mega-analysis using all the mice and the dense SNP set; this analysis identified several novel loci. Finally, we explored the replicability of associated loci discovered in either cohort by examining the same SNP in the other cohort. We have made these data publicly available on www.genenetwork.org, providing the community with a unique reference dataset.Author contributionsXZ imputed genotypes, performed SNP and individual QC, and conducted GWAS in F34 and F39-43 AILs under supervision of AAP. AAP also provided computational resources for the analyses in this paper. CS prepared GBS libraries for sequencing, as well as organizing portions of the F39-43 phenotypes. NMG de-multiplexed GBS sequencing results and performed alignment and variant calling. RC helped with kinship relatedness matrix calculated from AIL pedigree. AC provided technical support for running programs and scripts. GS collected F39-43 phenotypes, respectively. XZ co-wrote the manuscript with AAP, who designed the study and oversaw data collection.