Glutamic proteases (GPs) represent one of the seven peptidase families described in the MEROPS database of peptidases (also known as proteases, proteinases, and proteolytic enzymes). Currently, the GP family is divided into six sub-families (G1–G6) distributed across three clans (GA, GB, and GC). A glutamic acid and another variable amino acid are the catalytic residues in this family. Members of the GP family are involved in a wide variety of biological functions. For example, they act as bacterial and plant pathogens, and are involved in cancer and celiac disease. These enzymes are considered potential drug targets given their crucial roles in numerous biological processes. Characterizing GPs provides insights into their structure–function relationships, enabling the design of specific inhibitors or modulators. Such advancements directly contribute to drug discovery by identifying novel therapeutic targets and guiding the development of potent and selective drugs for various diseases, including cancers and autoimmune disorders. To address the challenges associated with labor-intensive experimental methods, we developed GPpred, an innovative support vector machine (SVM)-based predictor to identify GPs from their primary sequences. The workflow involves systematically extracting six distinct feature sets from primary sequences, and optimization using a recursive feature elimination (RFE) algorithm to identify the most informative hybrid encodings. These optimized encodings were then used to evaluate multiple machine learning classifiers, including K-Nearest Neighbors (KNNs), Random Forest (RF), Naïve Bayes (NB), and SVM. Among these, the SVM demonstrated a consistent performance, with an accuracy of 97% during the cross-validation and independent validation. Computational methods like GPpred accelerate this process by analyzing large datasets, predicting potential enzyme targets, and prioritizing candidates for experimental validation, thereby significantly reducing time and costs. GPpred will be a valuable tool for discovering GPs from large datasets, and facilitating drug discovery efforts by narrowing down viable therapeutic candidates.
