ProteinChat

- ProteinChat is a versatile, multi-modal large language model designed to predict protein functions from amino acid sequences.- ProteinChat works in a similar way as ChatGPT. Users upload a protein sequence and ask various questions about this protein. ProteinChat will answer these questions in a multi-turn, interactive manner. - The ProteinChat system consists of a protein encoder, a large language model (LLM), and an adaptor. The protein encoder takes a protein sequence as input and learns a representation for this protein. The adaptor transforms the protein representation produced by the protein encoder into another representation that is acceptable to the LLM. The LLM takes the representation transformed by the adaptor and users' questions about this protein as inputs and generates answers. All these components are trained end-to-end. We use xTrimoPGLM-1B as the protein encoder.- To train ProteinChat, we designed (protein, prompt, answer) triplets from the functions and keywords from Swiss-Prot dataset, resulting in ~500k proteins and 1.5 million triplets.