2.2. Prediction of Basic Biochemical Characteristics of Erythrocytic Stage Antigens

Some of the fundamental biochemical properties of the selected proteins were predicted using different web servers. For antigenicity, VaxiJen v2.0 (threshold: 0.45) possesses 78.0% prediction accuracy, based on physicochemical features of a protein (http://www.ddg-pharmfac.net/vaxijen/VaxiJen/VaxiJen.html) [32]. Also, ANTIGENpro, available at http://scratch.proteomics.ics.uci.edu/, is an alignment- and pathogen-free predictor of antigenicity using microarray data. Another important aspect of a vaccine candidate is lack of allergenicity, which was predicted using two web servers [33]. AllergenFP v1.0, available at https://ddg-pharmfac.net/AllergenFP/method.html, differentiates allergens from nonallergens with a descriptor fingerprint approach [34], whereas AllerTOP v2.0, available at https://www.ddg-pharmfac.net/AllerTOP/method.html, performs mostly through k-nearest neighbors (k NN) [35]. The solubility of each protein was predicted using Protein-Sol server (https://protein-sol.manchester.ac.uk/), based on available data for Escherichia coli protein solubility in a cell-free expression system [36].

Preliminary physicochemical properties of the erythrocytic stage proteins were predicted using ProtParam server, available at https://web.expasy.org/protparam/, which estimates the molecular weight (MW), positively and negatively charged residues, isoelectric point (pI), in vitro and in vivo estimated half-life, instability index, aliphatic index, and grand average of hydropathicity (GRAVY) [37]. Moreover, three online bioinformatics tools from DTU Health Tech Services server (Denmark), available at https://services.healthtech.dtu.dk/, including SignalP-6.0 [38], Deep TMHMM [39], and DeepLoc-2.0 [40] were used to characterize the putative signal peptide, transmembrane domain, and subcellular localization for each protein.