Anticancer Peptides Classification Using Kernel Sparse Representation Classifier

Cancer is one of the most challenging diseases because its complexity, variability, and diversity causes. It has been major research topics over past decades, yet it still poorly understood. To this end, multifaceted therapeutic frameworks are indispensable. Anticancer peptides (ACPs) promising treatment option, but their large-scale identification synthesis require reliable prediction methods, which a problem. In paper, we present an intuitive classification strategy that differs from traditional xmlns:xlink="http://www.w3.org/1999/xlink">black-box method based on well-known statistical theory xmlns:xlink="http://www.w3.org/1999/xlink">sparse-representation classification (SRC). Specifically, create over-complete dictionary matrices by embedding xmlns:xlink="http://www.w3.org/1999/xlink">composition K-spaced amino acid pairs (CKSAAP). Unlike SRC frameworks, use efficient xmlns:xlink="http://www.w3.org/1999/xlink">matching pursuit solver instead computationally expensive xmlns:xlink="http://www.w3.org/1999/xlink">basis in strategy. Furthermore, xmlns:xlink="http://www.w3.org/1999/xlink">kernel principal component analysis (KPCA) employed to cope with non-linearity dimension reduction feature space whereas xmlns:xlink="http://www.w3.org/1999/xlink">synthetic minority oversampling technique (SMOTE) used balance dictionary. The proposed evaluated two benchmark datasets for parameters found outperform existing methods. results show highest sensitivity balanced accuracy, might be beneficial understanding structural chemical aspects developing new ACPs. Google-Colab implementation available GitHub page (https://github.com/ehtisham-Fazal/ACP-Kernel-SRC).