Lata Sneh, Bhasin Manoj, Raghava Gajendra P S
Institute of Microbial Technology, Chandigarh, India.
Methods Mol Biol. 2007;409:201-15. doi: 10.1007/978-1-60327-118-9_14.
The machine learning techniques are playing a vital role in the field of immunoinformatics. In the past, a number of methods have been developed for predicting major histocompatibility complex (MHC)-binding peptides using machine learning techniques. These methods allow predicting MHC-binding peptides with high accuracy. In this chapter, we describe two machine learning technique-based methods, nHLAPred and MHC2Pred, developed for predicting MHC binders for class I and class II alleles, respectively. nHLAPred is a web server developed for predicting binders for 67 MHC class I alleles. This sever has two methods: ANNPred and ComPred. ComPred allows predicting binders for 67 MHC class I alleles, using the combined method [artificial neural network (ANN) and quantitative matrix] for 30 alleles and quantitative matrix-based method for 37 alleles. ANNPred allows prediction of binders for only 30 alleles purely based on the ANN. MHC2Pred is a support vector machine (SVM)-based method for prediction of promiscuous binders for 42 MHC class II alleles.
机器学习技术在免疫信息学领域发挥着至关重要的作用。过去，已经开发了许多使用机器学习技术预测主要组织相容性复合体（MHC）结合肽的方法。这些方法能够高精度地预测MHC结合肽。在本章中，我们描述了两种基于机器学习技术的方法，即nHLAPred和MHC2Pred，它们分别用于预测I类和II类等位基因的MHC结合物。nHLAPred是一个用于预测67种MHC I类等位基因结合物的网络服务器。该服务器有两种方法：ANNPred和ComPred。ComPred使用组合方法（人工神经网络（ANN）和定量矩阵）对30个等位基因预测结合物，并使用基于定量矩阵的方法对37个等位基因预测结合物。ANNPred仅基于人工神经网络对30个等位基因的结合物进行预测。MHC2Pred是一种基于支持向量机（SVM）的方法，用于预测42种MHC II类等位基因的混杂结合物。
