cACP-DeepGram

'cACP-DeepGram' is a software tool designed to address the challenges associated with cancer treatment using anticancer peptides (ACPs). Cancer is a significant global health concern characterized by irregular cellular growth and division, and traditional approaches to treat it, such as therapies and wet laboratory-based methods, can be costly and have diverse side effects. ACPs have garnered attention as a more selective and effective approach to cancer treatment, as they can target and treat cancer cells without harming normal cells.
The rapid increase in peptide sequences poses a challenge in accurately predicting the potential of ACPs for cancer treatment. 'cACP-DeepGram' aims to provide a reliable prediction model for ACPs. It employs a FastText-based word embedding strategy to represent each peptide sample through a skip-gram model. Once the peptide embedding descriptors are extracted, a deep neural network (DNN) model is applied to discriminate ACPs.
The optimized parameters of the DNN model achieved high prediction accuracy, with 96.94% accuracy in training samples, 93.41% accuracy in alternate samples, and 94.02% accuracy in independent samples.