结合基因组DNA功能元件和易感SNPs信息预测并鉴定骨质疏松症的特征性遗传标记

Osteoporosis is a common complex disease with high heritability, which can be prevented and controlled. The key issue for clinic prevention and treatment is to reveal its molecular genetic factors. Genome wide association studies (GWAS) is a well-accepted method of genetic epidemiology for identifying susceptibility genes on complex diseases. Although GWASs have identified a number of susceptibility genes for osteoporosis, they did not achieve the expected goal. The reasons might be two aspects. One is the limitation of the statistic power of GWAS. The other is the lack of knowledge of the function of non-coding sequence, while the most of variations identified by GWAS are located at non-coding regions. It is hard to explain the relationship between variations and disease. Recently, ENCODE published the fist-stage results. The whole genome sequences, especially for the non-coding regions, are detailed annotated by functional elements. It gives a clue to solve the two limitations mentioned above. Therefore, in order to solve above problems, this project will integrate the results of GWASs and ENCODE functional elements to extract characteristics of functional elements for osteoporosis by enrich analysis. Then, we will use these characteristics to predict new susceptibility variants for osteoporosis. These variants will be genotyped and validated in our samples. The validated variants will be used for constructing risk model for fractures. The results of this project will apply a new method and avenue for clinical early diagnosis, individualized prevention and treatment.

骨质疏松症是一类可防控的常见疾病，具有高遗传度，解决其预防和诊疗的根本在于结合临床揭示其分子遗传致病因素。尽管先进的全基因组关联研究（GWAS）发现一系列易感基因，效果却远未达预期。其原因一方面是GWAS的统计效力存在局限性；另一方面发现的变异多位于基因组非编码区，由于缺少对非编码区的认识，难以解释功能，从而忽略其重要性。近来，ENCODE恰逢其时地实施，对非编码区进行了功能元件详细注释，为疾病研究带来新的契机。本研究将延续前期工作，针对上述两个问题，创新性地将ENCODE数据和GWAS研究相结合，提取疾病易感SNPs的功能元件特征，建立全基因组水平预测疾病特征性遗传标记的方法；同时结合实际已有的临床大样本进行验证，高效鉴定影响中国人骨质疏松的特征性遗传标记；最后应用这些富含生物学功能的遗传标记并结合环境因素建立骨折风险预测模型，为疾病的临床早期筛查和个体化防治提供新思路和新靶点。

骨质疏松症是典型的多基因调控的复杂疾病，具有高遗传力。尽管全基因组关联分析发现了一系列骨质疏松相关基因，效果却远未达预期。其原因是由于发现的变异多位于基因组非编码区，难以解释功能，从而忽略其重要性。近年来，针对表观调控功能解析的研究如雨后春笋般兴起，为疾病非编码区变异的研究带来新的契机。本研究将表观调控数据和全基因组关联分析结果相结合，建立了预测富含生物学功能的疾病特征性遗传标记的方法，并利用新方法预测和验证骨质疏松症新的易感基因，提高了骨折风险预测模型效果。.首先，我们利用转录因子富集分析，发现2个显著富集于骨质疏松症相关增强子的转录因子EZH2(Padj=0.028)和NRSF(Padj=0.038)，且EZH2基因上的一个SNP rs111851041在多重检验校正后仍与中国人群的髋部和脊柱骨密度相关（髋部骨密度：P=4.32×10-4；脊柱骨密度：P=2.72×10-3），因此EZH2可作为骨质疏松症的易感基因。在人间充质干细胞成骨分化12~48小时，EZH2基因表达显著下调，证明了该基因与骨质疏松症相关。其次，我们进一步利用表观调控元件特征，建立了一种预测复杂疾病易感基因的方法，通过提取表观调控元件特征，从全局上反向预测复杂疾病易感基因。将该方法应用到骨质疏松上，发现预测的基因显著富集在已知的骨质疏松症相关通路上，如Wnt signaling, calcium signaling等。该方法已获得专利。最后，我们进一步利用反向预测排名靠前的基因构建多基因风险评分，并将该评分纳入原始骨折风险预测模型后，发现加入预测基因能显著提高骨折预测的准确性。.本研究创新性地将表观调控元件数据和全基因组关联分析结合起来，发现传统的全基因组关联研究无法检测到的遗传信号，为骨质疏松症的早期遗传筛查和防治提供新思路和新靶点。