Dissection Of The Genetic Structure Of C1sls Mouse Population And Identification Of The Candidate Genes In Blood Lipid Related Qtls

As one of the most important model organisms,mouse has been playing a significant role in the dissection of the human complex traits and diseases.Traditionally,two main strategies,linkage study and association study,were used to map the quantitative trait locus?QTL?and the underlying functional genes.However,the identified QTLs usually encompass large genomic intervals due to the fact that classical inbred strains originated from a handle of fancy mice which significantly limited their genetic diversity.Therefore,new mouse resources are needed.Chromosome substitution strains?CSSs?aim to increase the mapping power through decreasing the complexity of the genomic background.Recently,wild mice are increasingly becoming a promising mouse resource for QTL mapping because of its abundant genetic variations and accumulated recombination events which resulted in small haplotype blocks similar to those in human populations.Combining the respective advantages of CSSs and the wild mice resources,we proposed a new strategy to fine mapping and gene identification using a population of specific CSSs.Recently,we have constructed a population of chromosome 1?Chr 1?substitution lines?C1SLs?in which 30 mice were used as Chr 1 donors and C57BL/6J as the recipient strain.We have whole genome sequenced 18 lines of this population.In this doctoral dissertation,we focused on the dissection of the genetic structure through bioinformatic analysis.In addition,we evaluated the blood lipid traits diversity of C1 SLs mouse population and identified the candidate genes in known QTLs.the following are the main findings.First,Dissection of the genetic structure of C1 SLs.Compared with inbred strains,C1 SLs harbored abundant genetic variations.We totally identified 4.5 million single nucleotide polymorphisms?SNPs?and indels.Among them,1.3 million were classified as novel variants which did not exist in the Mouse Genome Project?MGP?sequenced 36 inbred strains and db SNP142 database.Functional annotation of the SNPs and indels identified 21335 variants located in protein coding regions which totally influenced 1002 genes.Among them,3009 missense variants,45 stop gains,6 stop losses,and 65 frameshifts were defined as novel variants.Further GO enrichment analysis revealed that 138 genes were involved in 60 known biological pathways including immune-related pathway,olfactory transduction.We also found 605 deletions in 205 genes caused loss of function,in which 3 have been related to human diseases and 7 genes were in 11 KEGG pathways.The genetic structure of C1 SLs is significantly different from laboratory mouse strains.In order to dissect the genetic structure of the C1 SLs population and the differences with inbred strains,we performed principal component analysis and found a significant difference in genetic structure between C1 SLs and MGP sequenced 36 inbred strains.Further subspecies origin analysis confirmed that the Chr 1 of each C1 SLs had a mosaic genomic structure which consisted of all the three subspecies' sequences.For B6-Chr1^KM,similar to the inbred strains,the Chr 1 sequence mainly derived from M.m.domesticus.For the other 17 C1 SLs whose Chr 1 came from Chinese wild mice,the Chr 1 sequence was mainly derived from M.m.musculus and M.m.castaneus.Their contribution was consistent with the geographical distribution of subspecies in China.That is,M.m.musculus mainly in the north while M.m.castaneus in the south.In addition,haplotype block analysis was performed according to the principle of the four gamete rule.Nearly 56% of blocks were < 1 kb in length,with only 4% ?10 kb,and the largest block was only 103.8 kb.We also found that C1 SLs had a higher nucleotide diversity than the MGP population.Second,identification of the selection pressure in wild mice and laboratory strains.It is generally assumed that laboratory mice may experience strong artificial selection while wild mice more natural selection.However,the differences of selection signature on the mouse genome and underlying genes between wild and laboratory mice remain unclear.In this study,we used two mouse populations: C1 SLs derived from Chinese wild mice and MGP sequenced inbred strains and two selection detection statistics: Fst and Tajima's D to identify the signature of selection on Chr 1.For Tajima's D,a total of 149 selection regions which encompass 7.215 Mb were detected in the C1 SLs.While for the MGP,we identified nearly twice selection regions?243?compared with the C1 SLs which accounted for 13.27 Mb Chr 1 sequence.We found very little overlaps of the selection regions between the C1 SLs and MGP,with only thirteen protein coding genes both identified in the C1 SLs and MGP.Moreover,the direction of selection is opposite in most of these genes?nine genes?.Through functional annotation,we identified several biological processes with significant enrichment including seven genes in the olfactory transduction pathway.For Fst,110 candidate selection regions containing 47 protein coding genes were detected.In addition,we searched the phenotypes associated with the 47 candidate selection genes.Results showed that most of them were involved in behavior,growth or body weight,mortality or aging,and immune systems which align well with the phenotypic differences between wild and laboratory mice.Third,Dissection of the genetic structure of B6-Chr1^KM mouse.The Chinese Kunming?KM?mouse is a widely used outbred mouse stock in China.However,its genetic structure remains unclear.Based on the results of the C1 SLs genetic structure,here we provide the detailed genetic variation and subspecies constitution of B6-Chr1^KM in which Chr 1 was derived from one KM mouse.With 36.6× average coverage of the entire genome,0.48 million SNPs and 96,679 indels were detected on Chr 1.Moreover,46,590 of them were classified as novel mutations.Further functional annotation identified 155 genes harboring potentially functional variants,among which 27 genes have been associated with 49 human diseases.We performed sequence similarity analysis using the SNPs identified on Chr 1 and their counterparts in three subspecies and found Chr1 of B6-Chr1^KM contained a large number of fragments with high sequence similarity to WSB?M.m.domesticus?.In addition,a bimodal distribution of blocks with two peaks of similarity was observed in a comparison of B6-Chr1^KM Chr 1 with PWK?M.m.musculus?counterpart.indicating that the sequence of M.m.musculus introgressed into the KM mouse.We next performed sliding window similarity analysis using 500 kb windows and 100 kb sliding intervals,found 13.5% and 6.4% of the Chr 1 sequences had high similarity?>99.7%?with PWK and CAST?M.m.castaneus?,respectively.To determine the extent of phylogenetic discordance in B6-Chr1^KM Chr 1,we assessed the discordance along Chr 1 with Bayesian Concordance Analysis.87.7% of the loci supported a single KM/WSB topology with higher posterior probability,and 9.7% supported a single KM/PWK topology.From both analyses,we can draw the conclusion that Chinese KM mouse has a mosaic genome structure with sequences predominately derived from M.m.domesticus and with at least some of the remaining sequences derived from M.m.musculus.Fourth,identification of the candidate genes in blood lipid QTLs.This study selected 13 lines of C1 SLs and one recipient strain C57BL/6J.By feeding high fat or control diet,we evaluated the total cholesterol?CHOL?,high-density lipoprotein cholesterol?HDL-C?,lowdensity lipoprotein cholesterol?LDL-C?,and triglyceride?TG?phenotypic diversity in this population.The results showed that there was a significant difference in the level of blood lipids between the C1 SLs mice and C57BL/6J.In addition,some lines were easy to induce and some of them were not.Several candidate genes were identified through differential gene expression analysis and genotype-phenotype association studies.First,3 control C1 SLs showing significant difference with C57BL/6J were selected to assay liver transcriptome expression levels and compared with the recipient.42 protein coding genes and 24 non-coding RNA were differentially expressed in the known QTLs.Among them,Spp1,Cd68,Cd74,Csf1 r,Acacb,Aldh1a7,and Akr1c19 genes had interaction with known lipid regulatory gene Soat1.Because CHOL,HDL-C,and LDL-C levels were significantly different from those of C57BL/6J in the three substitution mice.Gm70,Gm16432,Lyplal1,Actg-ps1,and Aldh1a7 with the same expression trends in the three lines were selected as candidate genes.For TG,expression levels of a protein coding gene?Acacb?and two unnamed non-coding RNA genes which were simultaneously declined in B6-Chr1^KM and B6-Chr1 SJ were identified.In addition,we performed genotype-phenotype association studies in the C1 SLs and 17 inbred strains separately.In the C1 SLs population,41 SNPs were significantly associated with the blood lipid traits while 69 were in the inbred strains.There are three hit genes present both in the C1 SLs population and the inbred population,namely the Cfap126 gene associated with CHOL,the 1700025G04 Rik gene associated with LDL-C and the TG-related Rbm44 gene.