论文部分内容阅读
目的建立一种基于近红外光谱-支持向量机(SVM)技术的单碱基差异识别方法。方法以仅相差1个碱基对的4种双链DNA为研究对象,其近红外光谱为识别变量,以径向基核函数(RBF)SVM建立非线性识别模型。结果对于长度为100bp的DNA链,当正则化系数γ=0.1,惩罚系数C=106时,模型的支持向量数最小为32,识别正确率为100%。结论该方法可发展为一种新的检测单核苷酸多态性的方法,具有简单、快速、低成本等优点。
Aim To establish a single base difference identification method based on near infrared spectroscopy and support vector machine (SVM). Methods Four kinds of double-stranded DNA, which differ only by 1 base pair, were selected as research objects. The near-infrared spectra were used as identification variables to establish a nonlinear identification model with Radial Basis Function (RBF) SVM. Results For the 100 bp DNA strand, when the regularization coefficient γ = 0.1 and the penalty coefficient C = 106, the minimum number of support vectors in the model is 32, and the correct recognition rate is 100%. Conclusion This method can be developed as a new method to detect single nucleotide polymorphism, which has the advantages of simple, rapid, low cost and so on.