Human papillomavirus (HPV) is recognized as the causative agent of cervical cancer in women, and it is associated with other anogenital and head/neck cancers. More than 120 types of HPV have been identified and many classified as high- or low-risk according to their oncogenic potential. One of its proteins, E6, has evolved to overcome the oncosuppressor functions of p53 by targeting this protein for degradation via interaction with the human ubiquitin-ligase E6AP. This study evaluates the correlation between the association strength of 40 HPV E6 types to the E6AP/p53 complex and the HPV oncogenesis risk using molecular simulations and machine and deep learning (ML/DL). In addition, a ML/DL-driven prediction is proposed for the HPV unclassified oncogenic risk type. The results indicate that thermodynamics play a pivotal role in the establishment of HPV-associated cancer and highlight the need to include some viral types in the HPV-related cancer surveillance and prevention strategies.
Keywords: binding free energy; cervical cancer; human papillomavirus; machine learning; molecular dynamics.
© 2022 Wiley Periodicals LLC.