The calculation of HSE is found by dividing a contact number (CN) sphere in two halves by the plane perpendicular to the Cβ-Cα vector.
If only Cα atoms are available (as is the case for many simplified representations of protein structure), a related measure, called HSEα, can be used.
[2] Song et al. have developed an online webserver termed HSEpred to predict half-sphere exposure from protein primary sequences.
[3] HSEpred server can achieve the correlation coefficients of 0.72 and 0.68 between the predicted and observed HSE-up and HSE-down measures, respectively, when evaluated on a well-prepared non-homologous protein structure dataset.
Recently, Heffernan et al. has developed the most accurate predictor for both HSEα and HSEβ based on a big dataset by using multiple-step iterative deep neural-network learning.