Identification of Ca2+-binding residues of a protein from its primary sequence

Abstract

ABSTRACT. Calcium is one of the most abundant minerals in the human body, playing a critical role in many cellular activities by interacting with different calcium ion (Ca 2+ )-binding proteins. Therefore, the correct identification of Ca 2+ -binding residues is essential for protein functional research. In this study, a new method was developed to predict Ca 2+ -binding residues from the primary sequence without using three-dimensional information. Through statistical analysis, four kinds of feature parameters were extracted from amino acid sequences: the increment of diversity values of amino acid composition, the matrix scoring values of position conservation, the autocross covariance of physicochemical properties, and the center motif. These features served as input for a support vector machine to predict Ca 2+ -binding residues. This method was tested on four well-established datasets using a five-fold cross-validation. The accuracies and Matthews correlation coefficients were 75.9% and 0.53 (dataset 1), 79.2% and 0.58 (dataset 2), 77.4% and 0.55 (dataset 3), and 79.1% and 0.58 (dataset 4). Comparative results show that the developed method outperforms previous methods. Based on this study, a web server was developed for predicting Ca 2+ -binding

Identification of metal ion binding sites based on amino acid sequences

“…If the central residue of the segment was a metal ion binding residue, then we assigned the segment as positive; otherwise, it was assigned as a negative segment. To generate the segment corresponding to the terminal residues in a protein sequence, we added an (L-1)/2 dummy residue "X" at both terminals of the proteins [ 30 – 34 ].…”

Recognizing Ion Ligand–Binding Residues by Random Forest Algorithm Based on Optimized Dihedral Angle

“…Among the feature parameters used to predict ion ligand-binding residues, position conservation information, and composition of amino acids were two commonly used basic feature parameters (Sodhi et al, 2004;Komiyama et al, 2015;Hu et al, 2016a;Liu et al, 2019;Wang et al, 2019). Besides, based on the biological background of interaction between ion ligands and proteins, researchers added physicochemical properties of amino acids, secondary structure, and relative solvent accessibility (RSA) to identify ion ligand-binding residues (Lin et al, 2006;Jiang et al, 2016;Cao et al, 2017;Li et al, 2017). Using these features obtained improved prediction results.…”

“…Thus, we constructed the position weight matrices to extract the 2L-dimensional position conservation information of amino acids. In terms of extraction of the hydrophilic-hydrophobic information of amino acids, autocross covariance formula was attempted (Jiang et al, 2016). However, the method did not take into account the different number of amino acid species contained in each class of hydrophilic-hydrophobic properties.…”

The Identification of Metal Ion Ligand-Binding Residues by Adding the Reclassified Relative Solvent Accessibility

“…Thus, depending on the interaction between the metal ion ligands and specific binding residues, many metal ion ligands can affect the special protein functions (Caspers et al, 1990;Supek et al, 1997;Selvarengan and Kolandaivel, 2005). For instance, Mn 2+ is used as catalyst in photosynthesis (Degtyarenko, 2000;Reed and Poyner, 2000), Ca 2+ can lead to anxiety and Alzheimer's disease (Jiang et al, 2015;Cao et al, 2017), and Cu 2+ can cause Coronary Heart Disease (Sodhi et al, 2004;Lin et al, 2005). The basic principle of molecular drug design is that the interaction between the receptor and ligand must conform to the "Lock and Key Model, " and the interaction between the protein and ion ligands we studied also conforms to the "Lock and Key Model."…”

“…Second, the feature parameters generally contained the composition information of the amino acid (Cao et al, 2017;Wang et al, 2019), hydrophilicity-hydrophobicity (Lin et al, 2005;Lin et al, 2006;Cao et al, 2017), charge (Lin et al, 2005;Cao et al, 2017;Wang et al, 2019), position specific score matrix (PSSM) (Hu et al, 2016a), relative solvent accessibility (RSA) (Lin et al, 2006;Hu et al, 2016a;Cao et al, 2017;Wang et al, 2019) and three-dimensional structure information (Babor et al, 2010;Roy et al, 2012;Yang et al, 2015;Hu et al, 2016a). Finally, the classification algorithms used were artificial neural network (ANN) (Lin et al, 2005), Support Vector Machine (SVM) (Lin et al, 2006;Jiang et al, 2015;Cao et al, 2017;Hu et al, 2016a), Naïve Bayes (Ebert and Altman, 2010), COFACTOR (Lin et al, 2006;Yang et al, 2015), TargetSeq, TargetCom (Hu et al, 2016b), COACH (Yang et al, 2015), and SMO (Wang et al, 2019). Among the three aspects in the prediction mentioned above, the key step of feature extraction was generated by one of two ways: (1) the three-dimensional structure information or (2) primary sequence information of the protein.…”