芋螺毒素是一种具有丰富序列多样性且富含二硫键的多肽.因其能够特异性地结合离子通道,阻断神经冲动的传递,芋螺毒索又具有极大的潜在药用价值.面对庞大的芋螺毒素种类,需借助理论的手段和方法来进行研究,从而降低实验成本,为实验做出理论指导.本文基于二项式分布结合支持向量机算法对芋螺毒素离子通道活性类型进行了预测,jackknife交叉验证的总体精度和平均精度分别达到87.5%和86.8%.结果表明,本文所建模型能够用于芋螺毒素的离子通道活性预测,为基于芋螺毒素新药物的进一步开发提供帮助和理论依据.
Conotoxins are small disulfide-rich peptide toxins and have an exceptional di- versity of sequences. Because they are able to specifically bind to ion channels and interfere with neurotransmission, they have become excellent pharmacological candidates in drug design. In order to save experimental cost and time, it's necessary to theoretically analyze the possible func- tions of newly sequenced conotoxins. In this paper, a novel method based on binomial distribution and support vector machine (SVM) was developed to predict the types of ion-channel targeted conotoxins. Our method achieved an overall accuracy of 87.5% with an average accuracy of 86.8% for three types of ion channel-targeted conotoxins classification in the jackknife cross-validation test, which are superior to that of other state-of-the-art classifiers. These results suggest that the proposed method will provide important instructions for theoretical and experimental research on conotoxins.