基于聚类算法的结构变异及其形成机制识别

在人类基因组上存在着涉及到不同序列长度的结构变异，这些结构变异对癌症的发生和发展产生了显著的影响。随着新一代测序技术的发展以及测序成本的降低使得在全基因组水平研究结构变异变得可能，基于聚类算法对千人基因组三个不同地区的样本以及ＣＧＨｕｂ数据库中结直肠癌样本进行了结构变异识别，并基于间断点处的序列同源性对结构变异的形成机制进行了分析；利用方差分析及非参数检验分析了结构变异和癌症的关系以及结构变异与地域之间的关系。最后，探讨了该领域未来的发展趋势。

Structural variation and its mechanism detection based on clustering algorithm

There are a number of structural variations of different lengths in the human genome, which has a significant impact on the occurrence of cancer. With the development of new generation high-throughput sequencing technology and the falling cost of sequencing make it possible to study structural variation at whole genome level. The samples from three different regions of the 1000 genomes and colorectal cancer samples from the CGHub database based on the clustering algorithm were analyzed in this paper and the mechanisms of the structural variation according to the sequence homology at the breakpoints was analyzed. The relationship between structural variation and cancer, as well as the relationship between structural variation and regional factors were analyzed by variance analysis and non-parameter tests. In the end, the future development in this research area was pointed out.