Toward predicting protein topology: An approach to identifying β hairpins

Abstract

Although secondary structure prediction methods have recently improved, progress from secondary to tertiary structure prediction has been limited. A promising but largely unexplored route to this goal is to predict structure motifs from secondary structure knowledge. Here we present a novel method for the recognition of ␤ hairpins that combines secondary structure predictions and threading methods by using a database search and a neural network approach. The method successfully predicts 48 and 77%, respectively, of all of hairpin and nonhairpin ␤-coil-␤ motifs in a protein database. We find that the main contributors to motif recognition are predicted accessibility and turn propensities.

The Comparison between Random Forest and Support Vector Machine Algorithm for Predicting β-Hairpin Motifs in Proteins

“…Because of the fixed-length pattern are generated using three rules, arbitrary sequence segments can obtain 6 S values (S12 (A)) which be calculated by Equation (4). f) Predicted secondary structure information (SS) In the research of predicting β-hairpin motifs, literature [2,3] had used predicted secondary structure information as the characteristic parameters; better prediction results were obtained. In order to improve the prediction effect, we also extract predicted secondary structure information.…”

“…In order to improve the prediction effect, we also extract predicted secondary structure information. These are obtained by using the PHD [2] software, and are represented by a vector of 3 dimensions which are the frequency of predicted secondary structure (α-helix, β-sheet and coils).…”

“…In the β-β motif, if two anti-parallel β-strands are connected by loop and there are one or more hydrogen bonds between two adjacent strands, then the structure is called as β-hairpin [1][2][3], otherwise it is considered as non-β-hairpin. Correct prediction of β-hairpin motifs is helpful to folding recognition, and it is vital for simplifying folding numbers of unknown structure [4][5][6].…”

“…In 2002, Cruz et al [2] employed an artificial neural network (ANN) for predicting β-hairpins in 534 protein chains; an accuracy of 47.7% was obtained. In 2004, Kuhn et al [1] also used ANN for predicting hairpins in 2209 protein chains by identifying local hairpins and non-local diverging turns; an accuracy of 75.9% was achieved.…”

Prediction of β-Hairpins in Proteins Using Physicochemical Properties and Structure Information

“…Elements of secondary structure and supersecondary structure can then combine to form the full threedimensional fold of a protein. As the information gained can be used in the tertiary structure prediction, protein supersecondary structure prediction is a key step in the hierarchical approach to derive the protein tertiary structure [9]. The protein supersecondary structure prediction methods developed thus far can be divided into two categories [1]: the first focuses on finding a number of different supersecondary structures at once [10], whereas the second is based on predicting special structures such as -hairpins.…”