| A run of homozygosity（ROH） defines a continuous or uninterrupted stretch of aDNA sequence without heterozygosity in the diploid state, that is in the presence ofboth copies of the homologous DNA segment. The phenomenon was first checked outin genome of tumor tissues or of Mendel diseases patients from consanguineousfamily. Long stretch ROHs were usually considered as driving factor of diseases,based on double-hit theory （stimulation of double recessive loci）. In a long time, thelack of method to detect genome-wide ROHs prevents researchers giving acomprehensive description of ROHs.Single nucleotide polymorhphisms （SNPs） is one of the most common geneticsvairants in the human genome, which defines single nucleotide variants withfrequency less than1%in the population, including conversion, transversion, deletionand insertion. Genotyping formation of sufficient SNPs in a region can reflect thewhole homozygosity levels in the region; thus, SNPs can be used to capture ROHs.Recently, genome-wide association based on SNPs array has been widely used toexplore genetics mechanism of complex diseases. Genome wide association studiesbased on SNP array provides genotyping information of tens of thousands SNPs,through which, researchers could detect ROHs in the scale of genome wide. Since2006, several groups explored and described ROHs in the data of HAPMAP project,concluding that ROHs was widely distributed in the human genomes. After that, someother researchers designed case-control study according to the hypothesis that ROHsmay cause the development of diseases by carrying low penetrance recessivediseases-related loci, and found the association between ROHs and multiple diseaseslike schizophrenia, Alzheimer’s disease and several cancers. However, few studiesdiscussed the ROHs distribution and their relationship with diseases in Chinese Hanpopulation.Since2012, our group started the first genome wide association study of non-cardia gastric cancer in Chinese Han. Based on the previous work, we designed agenome wide ROHs association study to describe ROHs levels in Chinese Han,explored the association between ROHs and non-cardia gastric cancer, and identifiednon-cardia gastric cancer related ROHs regions.The project included two independent studies, including565cases and1,162controls from Nanjing,468cases and1,123controls from Beijing. AffymetrixGeneChip Human Mapping6.0platform was used for GWAS genotyping. SNPs withlow call rate, low MAF, and deviated from HWE and individuals with low call rate,ambiguous gender, and potential relatives were excluded in the process of qualitycontrol. Health control subjects were used to evaluate overall ROH level. Logisticregression under additive model was used to assess the association between ROHsand non-cardia gastric cancer. FROH, which indicated the proportion of the autosomalgenome in runs of homozygosity, was used to reflect the overall ROHs level. T testwas used to detect difference between ROHs frequency of RegulomeDB annotatedregulation/function SNPs and non-regulation/function SNPs after log transformingROHs frequency of each SNP.In the outbred health control, ROHs over0.5Mb was widely spread in thegenome wide scale （Nanjing: FROH=5.08%, Beijing: FROH=4.95%）. According to thearticle of Pemberton et al., we classified ROHs into three classes （A:500Kb-689.346Kb; B:689.346Kb-1548.887Kb; C: over1548.887Kb）. Class A and Bwere moderately correlated （r=0.71, P<10-16）, but both of them were in weakcorrelation with Class C FROH（r=0.05for both classes A and B）. Thus in furtherassociation analysis, we combined first two classes （Class A&B）. In addition, wefound no significant relationship between age and ROHs （Nanjing: Class A&B:r=-0.015, P=0.51; Class C: r=0.010, P=0.67; Beijing: Class A&B: r=-0.038, P=0.27;Class C: r=0.020, P=0.56）, and more ROHs frequency in RegulomeDB annotatedregulation/function regions (Nanjing: Class A&B: P=2.82×10-12;Class C:P=6.97×10-6; Beijing: Class A&B: P=1.48×10-6; Class C: P=3.27×10-3), indicatingthe importance of ROHs in human genome.In the association analysis of ROHs, FROHof ROHs in Class A&B were significant associated with decreased risk of non-cardia gastric cancer in both Nanjingand Beijing studies (Nanjing: OR=0.56,95%CI:0.43-0.73, P=1.65×10-5; and Beijing:OR=0.66,95%CI:0.50-0.87, P=3.75×10-3, respectively), while effect of FROHinClass C ROHs in Nanjing study failed to replicate in Beijing （Nanjing: OR=1.14,95%CI:1.02-1.28, P=0.03; Beijing: OR=0.87,95%CI:0.72-1.06, P=0.18）. Tofurther locate non-cardia cancer related ROHs region, we scanned the ROHs regionwith frequency over5%. Among them, one ROH pool at15q13.2with the key SNP ofrs16956610was consistently associated with non-cardia gastric cancer risk in bothNanjing and Beijing studies (OR=0.55,95%CI:0.39-0.76, P=3.39×10-4; andOR=0.41,95%CI:0.24-0.71, P=1.55×10-3, respectively; Table2). After combiningthese two studies using meta-analysis, the association (OR=0.51,95%CI:0.38-0.67,P=2.49×10-6) reached genome-wide significance level after Bonferroni correction（0.05/16726）. In addition, another ROH pool at2q14.1with the key SNP ofrs17045361also showed consistent association with non-cardia gastric cancer inNanjing and Beijing studies (OR=1.69,95%CI:1.23-2.31, P=1.11×10-3; andOR=1.67,95%CI:1.04-2.65, P=3.00×10-2, respectively; Table2), though did notreach genome-wide significance level (OR=1.68,95%CI:1.30-2.18, P=9.06×10-5).In summary, our results find the importance of Class A&B ROHs and indicatethat the ROHs are low in non-cardia gastric cancer and may be implicated with thesusceptibility to non-cardia gastric cancer. We also identified two ROHs at15q13.2and2q14.1associated with risk of non-cardia gastric cancer using a genome-wideROH association analysis in Han Chinese population, which may be important tofurther elucidate the genetic determinants of non-cardia gastric cancer. However, theexact mechanisms through which these two regions play in the development ofnon-cardia gastric cancer are still unclear and need to be further investigated.Nerveless, this study represents the first effort in exploring the role of ROHs on aspecific disease in a genome-wide scale. |
PDF Full Text Request