基于实例的基因分类及确定基因标签模型

Models of gene classification and determining gene tag based on example

所建立的模型及所得的结论有利于利用数据库中已有的基因信息快速筛选出潜在的癌症相关基因,模型一和模型二以基因表达水平限值和差异显著性水平为分类要素,将基因分为两类.模型三利用逐步优化思想建立优化模型,确定出六组基因标签.模型四利用小波分析法去噪及相关性检验法,重新确定基因标签,包含8种特征基因,对癌症样本的检测率降低了,说明数据中的噪声能对确定基因标签产生有利的影响.

The models and conclusions in this paper which facilitate the usage of the existing gene infomation in the database to screen out the potential tumor-related genes rapidly are studied. The limit value of gene expression and the level of significance of difference are considered as the classification factors in modelⅠ and model Ⅱ, and genes can be divided into two categories. By using zhe stepwise optimizing idea, optimization model and six groups of optimal gene tags are obtained in model Ⅲ. By using the wavelet analysis and correlation test, the gene tags are re-determined in modle Ⅳ, including eight kinds of characteristic genes. By the decrease of the detection rate of cancer samples, that noise in the data produces beneficial effects for determining the gene tags is indicated.