Multi-class protein fold recognition using support vector machines and neural networks

Abstract

Most current discriminative methods for protein fold prediction use the one-against-others method, which has the well-known 'False Positives' problem. We investigated two new methods: the unique one-against-others and the all-against-all methods. Both improve prediction accuracy by 14-110% on a dataset containing 27 SCOP folds. We used the Support Vector Machine (SVM) and the Neural Network (NN) learning methods as base classifiers. SVMs converges fast and leads to high accuracy. When scores of multiple parameter datasets are combined, majority voting reduces noise and increases recognition accuracy. We examined many issues involved with large number of classes, including dependencies of prediction accuracy on the number of folds and on the number of representatives in a fold. Overall, recognition systems achieve 56% fold prediction accuracy on a protein test dataset, where most of the proteins have below 25% sequence identity with the proteins used in training.

Protein classification using texture descriptors extracted from the protein backbone image

“…The fold database used in our experiments is derived from the work of (Ding and Dubchak, 2001). It contains a training set and a testing set that contain 313 and 385 proteins.…”

“…In order to validate the effectiveness of the presented method, we compare it with several methods in different literatures all of which used the same benchmark data sets and the same testing protocol of original paper (Ding & Dubchak, 2001), it is the dataset named FOLD in this work. The comparison is listed in Table 4.…”

“…The comparison is listed in Table 4. (Ding & Dubchak, 2001) proposed six kinds of features denoted by 22 C,S,H,P,V and Z respectively. Letter C is just the popular amino acid composition, while the left five letters indicate the features of Polarity, Polarizability, Normalized Van Der Waals volume, Hydrophobicity and Predicted secondary structure respectively.…”

“…Method Accuracy (%) (Ding & Dubchak, 2001) 56.50 (Chinnasamy et al, 2005) 58.18 (Shi et al, 2006) 61.04 (Huang et al, 2003) 65.50 (Lin et al, 2007) 69.60 (Shi & Zhang, 2009) 72.99 FUS3 81.04 Table 4. Other comparisons in the FOLD dataset.…”

Support Vector Machines for Diagnosis of Breast Tumors on US Images

“…Recent research, however, has suggested that the SVM is superior to the neural network (13)(14)(15). To verify this, and to test our method, we used a multilayer propagation (MLP) neural network model, as well as an SVM, to classify tumors in our experiments.…”