Soft Computing Methods for Disulfide Connectivity Prediction

Abstract

The problem of protein structure prediction (PSP) is one of the main challenges in structural bioinformatics. To tackle this problem, PSP can be divided into several subproblems. One of these subproblems is the prediction of disulfide bonds. The disulfide connectivity prediction problem consists in identifying which nonadjacent cysteines would be cross-linked from all possible candidates. Determining the disulfide bond connectivity between the cysteines of a protein is desirable as a previous step of the 3D PSP, as the protein conformational search space is highly reduced. The most representative soft computing approaches for the disulfide bonds connectivity prediction problem of the last decade are summarized in this paper. Certain aspects, such as the different methodologies based on soft computing approaches (artificial neural network or support vector machine) or features of the algorithms, are used for the classification of these methods.

Residue Adjacency Matrix Based Feature Engineering for Predicting Cysteine Reactivity in Proteins

“…A ratio higher than one indicates that the amino acid substitution is favored. Ratios less than one indicate that that the amino acid substitution is not favored [ 32 ]. For a window size of 2*k + 1 (k positions to the left of the target cysteine, k to the right, and the target cysteine position itself) and the twenty major amino acids, we get a matrix that is twenty rows long and 2*k + 1 columns wide that can be vectorized for a total of 20*(2*k + 1) features.…”

Expanding the structural atlas of peptide toxins

“…103,[199][200][201][202][203] These methods, recently reviewed by Márquez-Charmorro et al, (2015), have a range of limitations. 204 The majority of connectivity prediction methods extract cross linking information from homologous proteins. Such approaches cannot be used to analyse peptides of uncharacterised peptide classes.…”

“…Other methods have predominantly used sequence and evolutionary information and are often reliant on finding high sequence matches. 254 CrossLink achieved a baseline accuracy of 82% with >90% accuracy for high probability predictions. A number of examples demonstrated how CrossLink can be a valid approach that can either predict the native connectivity outright or dramatically reduce the number of possible connectivity frameworks.…”

“…Existing methods to predict the disulfide connectivity in proteins predominantly use sequence and evolutionary information. 204 These methods often predict Cys connectivity by identifying homologous proteins for which connectivity has been validated experimentally. The benchmark datasets that were used to train these methods comprise larger proteins and few or no peptides.…”

“…The performance of over 20 Cys connectivity predictive methods that predominantly rely on evolutionary and sequence information have been benchmarked, using as testing sets data from Swiss-Prot release SP39 or from PDBCYS. 204,214,215 These two datasets do not contain peptides or proteins studied by NMR spectroscopy, therefore preventing a direct comparison with CrossLink results. We instead compared the performance of three publicly available connectivity programs, namely Disulfind, EDBCP and DiANNA, for the prediction of the Cys connectivity of the disulfide-rich peptides listed in Table 4-5.…”