All-atom diffusion-based generative model developed for de novo design of protein and peptide binders across a wide array of target types
A flexible de novo binder design workflow using RFdiffusion for backbone generation
Protein binder design pipeline that can be used to design miniprotein and peptide binders
Open-source framework for combining and composing a variety of structure prediction and other models for design, filtering, or guidance.
Fine-tuning pipeline combining reinforcement learning algorithms (GRPO, DPO) with the ZymCTRL protein language model.
pLM trained with masked diffusion to enable both high-quality representation learning and generative protein design
ESM2 is a masked language model trained on UniRef protein sequences. Used for structure prediction, property prediction, functional annotation, and clustering similar proteins.
PXDesign is a model suite for de novo protein-binder design — a diffusion generator (PXDesign-d) paired with Protenix and AF2-IG confidence models for selection. Across seven targets, PXDesign delivers 17-82% nanomolar hits on six. https://…
Motif-scaffolding pipeline that designs protein sequences around a given motif using the EvoDiff model.
BLOSUM (BLOcks SUbstitution Matrix) matrix is a substitution matrix used for sequence alignment of proteins
A reasoning model that proposes modifications to improve protein stability.
Antibody-specific transformer language model pre-trained on 558M natural antibody sequences.
ProTrek is a tri-modal protein language model that jointly models protein sequence, structure and function (SSF)
A workflow for designing protein binders using co-TIMED, a method for co-designing sequences in a protein complex
Foundry provides tooling and infrastructure for using and training all classes of models for protein design, including design (RFD3), inverse folding (ProteinMPNN) and protein folding (RF3). https://github.com/RosettaCommons/foundry
Llama 3 is a family of multilingual large language models (LLMs) which are pretrained and instruction tuned generative models
PPIFlow demonstrated a flow-matching-based framework and integrated design workflow for the de novo generation of high-affinity binders targeting precise epitopes. It supports a diverse range of tasks including protein-protein interaction (…
Extended RFdiffusion that incorporates cyclic relative positional encodings to enable the generation of the macrocyclic peptide backbones. https://www.nature.com/articles/s41589-025-01929-w https://github.com/RosettaCommons/RFdiffusion/blob…
