基于氨基酸组分和支持向量机的动物毒素的预测

Prediction of Animal Toxins Using Amino Acid Composition and Support Vector Machine

动物毒素是一种具有非常广泛用途的生物毒素,有必要提出一种能够快速、准确预测动物毒素的理论算法.基于动物毒素蛋白质序列的二肽组分信息,提出了一种离散增量结合支持向量机的ID-SVM的算法,并使用此算法对不同序列相似性的动物毒素进行了预测,取得了较好的结果.为了说明ID-SVM算法在预测生物毒素方面的优越性,在这里将ID-SVM算法应用到Saha和R aghava构建的细菌毒素和非毒素的数据库上,预测结果显示ID-SVM算法的预测结果高于Saha和R aghava所用算法的结果.

Animal toxins have a very important application in basic research. So,it is very important to predict them by using a computer method. Based on the 2-peptide components of local amino acid sequence,a novel ID-SVM algorithm combined increment of diversity （ID） with support vector machines （SVM） is proposed to predict animal toxins with different sequence identities; In order to estimate the effectiveness of this new algorithm, the bacterial toxin and non-toxin datasets generated by Saha and Raghava are also predicted. The higher predictive success rates than the previous algorithms are obtained by the ID-SVM algorithm.