Protein Fold Recognition Using an Overlapping Segmentation Approach and a Mixture of Feature Extraction Models

Abstract

Protein Fold Recognition (PFR) is considered as a critical step towards the protein structure prediction problem. PFR has also a profound impact on protein function determination and drug design. Despite all the enhancements achieved by using pattern recognitionbased approaches in the protein fold recognition, it still remains unsolved and its prediction accuracy remains limited. In this study, we propose a new model based on the concept of mixture of physicochemical and evolutionary features. We then design and develop two novel overlapping segmented-based feature extraction methods. Our proposed methods capture more local and global discriminatory information than previously proposed approaches for this task. We investigate the impact of our novel approaches using the most promising attributes selected from a wide range of physicochemical-based attributes (117 attributes) which is also explored experimentally in this study. By using Support Vector Machine (SVM) our experimental results demonstrate a significant improvement (up to 5.7%) in the protein fold prediction accuracy compared to previously reported results found in the literature. 4 The experimental results achieved in this step for all five classifiers for EDD and TG benchmarks are available upon request. 5 Details about the attribute selection process as well as the list and references of all 117 physicochemical-based attributes are available upon request.

Gram-positive and Gram-negative protein subcellular localization by incorporating evolutionary-based descriptors into Chou׳s general PseAAC

“…Among all these methods, pseudo amino acids composition has attained the best results to extract local information Wan et al, 2013;Esmaeili et al, 2010;Chou, 2001Chou, , 2005Chou, , 2011. In the present study, we extend the concept of segmented distribution features as described in the previous subsection to compute the auto covariance features from the segmented protein sequence (Dehzangi et al, 2013d). This is done to enforce local discriminatory information extracted from PSSM.…”

“…We do not do this normalization which maintain the general total occurrence of the amino acids based on their substitution scores. Since we do not normalize by dividing it by length, it becomes implicitly a part of that feature (Taguchi and Gromiha, 2007;Sharma et al, 2013b;Dehzangi et al, 2013d).…”

SumSec: Accurate Prediction of Sumoylation Sites Using Predicted Secondary Structure

“…Thus, the bigram occurrence matrix B will consist of all the frequencies Bp,q. In this study, we have employed profile bigram because of its promising results [41,66,67,68,69,70,71,72,73]. Each bigram matrix B is then transformed to one feature vector as F=[B11,…,Bij,…,B3,3]T for i=1,2,3 and j=1,2,3, where superscript T denotes transpose.…”

EvoStruct-Sub: An accurate Gram-positive protein subcellular localization predictor using evolutionary and structural features

“…This feature was introduced in [5, 89,86]. It was shown that information about the interaction of neighboring amino acids along the protein sequence can play an important role in providing significant local discriminatory information and enhancing protein subcellular localization prediction accuracy [6,87].…”

“…This feature group is 145 directly extracted from PSSM matrix. It aims at capturing global discriminatory information regarding the substitution probabilities of the amino acids with respect to their positions in the protein sequence[85,86,87]. This feature is extracted by summation of the substitution score of a given amino acid with all the amino acids along the protein sequence.…”