染色体突变区间的识别算法研究

Mutation region detection is done via linkage analysis. Today linkage analysis serves as a way of identifying disease causal mutations. Linkage studies have facilitated the identification of several hundred human genes that can harbor mutations that by themselves lead to a disease phenotype. The fundamental problem in linkage analysis is to identify regions whose allele is shared by all or most affected members but by none or few unaffected family members based on the genotype data of a set of individuals. Several famous programs have been developed for linkage analysis when the pedigree within three generations is known and there are enough number of individuals in the pedigree. Those programs can give very accuracy outputs under many but not all cases. Very often, in a pedigree of three generations, there are few number of individuals whose genotype data are available. In this case, the existing methods do not work.To solve the problem, people try to include more individuals in a pedigree of 6 or more generations. The problem becomes mich harder in this case since the genotype data of individuals in the earlier generations cannot be obtained and the relationship between the individuals with genpotype data are far away. To our knowledge, no algorithm can give good outputs when the sampled individuals share common ancesters six generations ago. With the new development of microarray techniques, high-density SNP genotype data can be used for large-scale and cost-effective linkage analysis. Recently, the international HapMap project has produced enormous amount of haplotype data for individuals in some major populations. These new developments make it possible for us to propose new mathematical models for finding genes causing genetic diseases when the sampled individuals share common ancesters six generations ago. This project emphasizes algorithmic issues for all the proposed computational problems. We will also try to implement our designed algorithms to form aoftware packages that work well in practice.

在遗传学上,染色体突变区间的确定通常是通过连锁分析（linkage analysis）进行的。这一方法的应用，已经获知了几百种由自身的基因变异引起的显性疾病。连锁分析的方法是根据一组个体的基因型遗传标记（genotype data）来推断致病基因在染色体中所在的区间。若数量足够多的一组个体是在一个近三代家谱中，已经有有效的实用程序可以解决。在某些情况下，一个近三代的家谱中只能找到少数几个人的基因型标记数据。因此，现有方法无法解决此问题。解决的方法之一是利用更多个体的基因型标记。因此，我们要考虑具有六代或以上的共同祖先的一组个体。然而，在此情况下，问题就变得复杂。在这样大的家谱中，很多先人的基因型标记数据已无法获得，而个体之间的血缘关系变得疏远。目前，尚无任何算法可解决此问题。在此项目中，我们将根据国际人类基因组单体型图计划所得到的数据提出新的数学模型解决此问题。

我们按照项目计划书对染色体区间识别的算法进行了研究。作为该课题的扩展延伸，我们进行了关于单倍型组装问题的研究。另外，我们也对其他相关研究方向，例如，泛基因组比较的研究，蛋白质结构位点预测，阻塞模式匹配问题的研究及谱系比较的研究。， .共在国际学术期刊发表论文11篇。.有几个算法已转换成计算机软件。.我们还成功举办了COCOA2016 及APBC 2017两个国际会议。.此外，我们正在筹划Cocoon2018及CPM2018国际会议。将于2018年7月2日至7月4日于青岛同时举行。.我们参加了几个国际会议并发表了论文。.我们与日本东京电器大学（ZHENZHi-Zhong 教授），加拿大阿尔伯特大学计算机系(LIN Guohui教授, 澳大利亚悉尼科技大(LI Jinyan教授),山东大学(朱大铭教授)，天津大学（郭菲副教授），中南大学（王建新教授，李敏教授）合作并联合发表了论文。