ANNAVP uses various types of neural networks to facilitate the prediction of the efficacy of an already tested antibody to new strains. ANNAVP can use supervised or unsupervised learning and viral protein sequence data to form correlations between different strains and to predict the efficacy of neutralizing agents against them.
See Wiki: https://github.com/icghita/ANNAVP/wiki