NanoMelt: Semi-supervised ensemble model for nanobody thermostability prediction

NanoMelt is a semi-supervised ensemble model trained to predict the apparent melting temperatures of nanobodies. The model involves a Ridge stacking of multiple Gaussian Process Regressors (GPRs) and Support Vector Regressors (SVRs) combined with diverse pre-trained nanobody embeddings(i.e., one-hot, VHSE, ESM-1b, and ESM-2).