PoET#
Protein Evolutionary Transformer (PoET) is a machine learning model that generates and identifies protein sequence variants with improved function and stability. PoET requires a prompt, (ie. a set of sequences representing the target protein sequence distribution). This will commonly be an evolutionary context prompt in the form of filtered MSAs (Align). But PoET can accept any collection of sequences!
Endpoints to call our generative PoET model for de novo generation of proteins, evaluation of protein fitness, and single site mutant analysis of proteins. These workflows are all possible without prior wet lab data, and therefore do not require assaydata to be pre-loaded! The endpoints enable scoring and generating sequences from a prompt.