RosettaMPNN documentation¶
Welcome to the RosettaMPNN documentation! To contribute to the documentation, please refere to the contributors guide, linked below.
Overview¶
RosettaMPNN combines all of the capabilities of ProteinMPNN, LigandMPNN, and HyperMPNN into a single, Rosetta Commons-supported tool for protein sequence generation.
If you would like your MPNN-based tool incorporated into this repository, create a pull request or reach out to Hope Woods, the Rosetta Commons Technical Product Lead.
ProteinMPNN forms the basis for LigandMPNN, HyperMPNN, and now RosettaMPNN. It is a message-passing neural network that can generate protein sequences based on backbone structures. Due to its ability to couple amino acid sequences in different chains and awareness of symmetry, it can be used to design monomers, cyclic oligomers, protein nanoparticles, and protein-protein interfaces. We have also included the Multi-State design capabilities developed by the Kuhlman Lab to enable sequence design for multiple protein conformations.
LigandMPNN extends the capabilities of ProteinMPNN to also be able to design protein sequences in the context of small molecules, nucleotides and metals. This allows for the design of small molecule binding proteins, sensors, and enzymes.
HyperMPNN uses ProteinMPNN’s model architecture and training algorithm, but includes proteins found in hyperthermophilic organisms to generate highly thermostable proteins. These proteins are incredibly useful for the creation of vaccines, protein nanoparticles for drug delivery, and industrial biocatalysts.
RosettaMPNN can be used in a workflow to design de novo proteins that combine backbone generation tools - such as RFdiffusion - and 3D structure prediction tools - such as AlphaFold2.
Contents: