Abstract: The present study for locating disease genes by IBD mapping needs to detect IBD segments firstly, and then performs association analysis using the IBD information. But it takes a long time to detect IBD segments. A new algorithm FADG is proposed to find the candidate IBD segments by analyzing the relationship between IBS and IBD, and defines evaluation function to analyze case and control groups separately, the largest difference position of SNP between the two groups is the causative loci. The experimental analysis for effectiveness, based on the five major causative loci of rheumatoid arthritis provided by GAW15 and the data which generated by the program GS, shows the results are consistent with the answer. The algorithm shows higher efficiency in ensuring the effectiveness by comparing with existing method. Finally, the permutation test for two chromosomes from data RA verifies the credibility, and further excludes the possibility of a false positive result.

Key words: Genome Wide?Association Study（GWAS）, Identity By State（IBS）, Identity By Descent（IBD）, case-control study, Single Nucleotide Polymorphism（SNP）

摘要： 目前通过IBD定位研究致病基因首先需要检测IBD片段，然后利用得到的IBD关系进行关联检测来定位致病基因，但是寻找IBD片段需要较长时间。提出了一种新的算法FADG，通过分析IBS和IBD的关系找出候选IBD片段，然后定义评价函数分别对病例组和对照组进行分析，差异最大的SNP位点就是致病位点。通过对GAW15提供的类风湿关节炎模拟数据的五个主要的致病位点和GS生成的模拟数据进行分析，实验结果与数据给定的致病位点一致；与已有的方法进行对比表明该算法在能找到正确位点的基础上效率较高；最后选择了RA数据中两条染色体分别进行permutation测试，验证算法得到致病位点的可信度，进一步排除所得结果假阳性的可能性。

关键词: 全基因组关联研究, 状态一致（IBS）, 同源一致（IBD）, 病例对照研究, 单核苷酸多态性