前列腺癌细胞核酸适配体支持向量机分类模型  

Support Vector Classification of Aptamers for Human Prostate Cancer

作　　者：禹新良[1,2] YU Xinliang(Hunan Provincial Key Laboratory of Environmental Catalysis&Waste Regeneration,Hunan Institute of Engineering,Xiangtan 411104;State Key Laboratory of Chemo/Biosensing and Chemometrics,Hunan University,Changsha 410082)

机构地区：[1]湖南工程学院环境催化与废弃物再生化湖南省重点实验室 [2]湖南大学化学生物传感与计量学国家重点实验室

出　　处：《分析科学学报》2020年第1期19-25,共7页Journal of Analytical Science

基　　金：国家自然科学基金(No.21190040,21190041);湖南教育厅重点项目(No.16A047);化学生物传感与计量学国家重点实验室(湖南大学)开放课题(No.2016013)

摘　　要：本文将前列腺癌(Prostate Cancer,PCa)PC-3M-1E8细胞为标靶的核酸适配体序列翻译成氨基酸序列,计算氨基酸序列的分子参数,然后用这些分子参数建立核酸适配体亲和性的构-效关系模型。所用的候选核酸适配体序列是采用以细胞为靶标的指数富集配体系统进化(Cell-SELEX)技术筛选得到。模型训练集、测试集分别包含150、50条核酸序列,均由第3轮和11轮的候选核酸适配体组成。将第3轮的核酸序列类标签值设置为“1”,代表低亲和性、低特异性的候选核酸适配序列;将第11轮筛选所得核酸序列类标签值设置为“2”,代表高亲和性、高特异性候选核酸适配序列。基于二值分类问题的支持向量机分类(SVC)算法用于建模。SVC模型对训练集、测试集的预测准确度分别为87.3%、86%。另外,采用SVC模型对第5、7、9轮的序列也进行了预测。第3、5、7、9、11轮的高亲和性与高特异性核酸适配体的分率分别是0.23、0.41、0.61、0.64、0.87,预测结果符合SELEX筛选的适配体进行规律。To develop a classification model for affinity of aptamers against prostate cancer(PCa)cell line PC-3M-1E8,molecular descriptors for PCa aptamers were calculated from amino acid sequences that were translated from DNA aptamer sequences.Candidate aptamers used in paper were obtained with cell based systematic evolution of ligands by exponential enrichment(cell-SELEX).The training set consisted of 150 candidate aptamers from the 3rd and 11th rounds of cell-SELEX with PCa cell line PC-3M-1E8 as target cells;and the test set was comprised of 50 aptamers from the 3rd and 11th rounds.The class labels of candidate aptamers from the 3rd round were set as 1,denoting the low affinity and specificity aptamer candidates.The class labels from the 11th round of SELEX were defined as 2,denoting the high affinity and specificity aptamer candidates.Support vector classification(SVC)technique for a two-class problem was used to develop the model.The SVC classification model in this paper had prediction accuracy of 87.3% for the training set and 86.0% for the test set.In addition,the SVC model was used for the prediction of candidate aptamers from the 5th,7th,and 9th rounds of cell-SELEX.The predicted fractions of aptamers with high affinity and specificity that were obtained from the SELEX selection of the 3rd,5th,7th,9th and 11th rounds were 0.32,0.41,0.61,0.64,and 0.87,respectively,which conformed to the aptamer evolutionary principles of SELEX based screening.

关 键 词：氨基酸 核酸适配体 前列腺癌 支持向量机分类 

分 类 号：O6[理学—化学]