"Revolutionary Deep Learning Techniques for Antiviral Medicine Discovery and ICU Prediction in COVID-19 Patients"
Researchers have proposed a deep learning model called Deep-AVPiden, based on temporal convolutional networks (TCNs), for the classification and discovery of antiviral peptides (AVPs). The model outperforms existing classifiers and has a less compute and memory-intensive version called Deep-AVPiden (DS). A web app has been developed using both models to aid wet-lab researchers in discovering AVPs in protein sequences. The models were trained on a dataset of peptides collected from various sources and were compared using statistical analysis. The proposed models have shown promising results in identifying AVPs in proteins of plants, mammals, and fishes, with 15 AVPs proposed for chemical synthesis and experimental validation.