基于混合泊松分布的新生突变识别算法

发布时间：2018-05-11 20:26

本文选题：人类基因组 + 新生突变　；参考：《中国生物化学与分子生物学报》2017年11期

【摘要】：对个体而言,不经父母遗传而后天获得的突变称为新生突变,绝大多数癌症都起自新生突变。构建快速精确的变异识别算法将有助于对癌症的研究。然而,针对前期新生突变识别算法准确率不高,且耗时多等问题,本文引入了基于变异位点的先验概率分布模型,运用基于混合泊松分布的期望最大化(EM)算法对新生突变识别算法进行改进与优化,研究了有亲缘关系的新生突变的识别,并在识别精度与运算速度方面与已有算法进行对比。结果表明,基于混合泊松分布的期望最大化算法在提高运算速度的同时降低了假阳性比率,具有良好的识别效果。
[Abstract]:For individuals, mutations acquired without parental inheritance are called new mutations, and most cancers originate from new mutations. The construction of a fast and accurate mutation recognition algorithm will be helpful to the study of cancer. However, in order to solve the problems of low accuracy and time consuming, a priori probability distribution model based on mutation locus is introduced in this paper. Using the expectation maximization (EMM) algorithm based on the mixed Poisson distribution to improve and optimize the new mutation recognition algorithm, this paper studies the genetic relationship of the new mutation recognition, and compares the recognition accuracy and the operation speed with the existing algorithms. The results show that the expected maximization algorithm based on the mixed Poisson distribution can improve the operation speed and reduce the false positive ratio.
【作者单位】：北京科技大学数理学院信息与计算科学系;河南偃师高级中学;
【基金】：国家自然科学基金(No.61370131)资助~~
【分类号】：O211;Q987

【相似文献】