The approach we take here is purely based on sequence. We use a PLM to embed the sequence of the Nipah virus together with that of its known receptor. We then take the embeddings of the interface residues of the Nipah virus as our "reference embeddings". We then use an MC sampling approach ( using BAGEL (https://github.com/softnanolab/bagel) in sequence space to look for a sequence of aminoacids that, once embedded with that of the Nipah virus as done for the known receptor, would make the embedding of the interfacial residues as close as possible to the original reference embeddings. The hypothesis is that if the embeddings contain information about the local environment in the folded state, by recreating the embeddings in a known complex, we can obtain a binder.