De Novo Biomolecule Design
Antibody / biologics design
for antibody-antigen complex prediction (still a hard task).
diffusion models specialized for CDR loops and full antibody scaffolds.
antibody-specific language models.
for antibody-antigen complex prediction (still a hard task).
antibody-specific language models.
for antibody-antigen complex prediction (still a hard task).
diffusion models specialized for CDR loops and full antibody scaffolds.
antibody-specific language models.
Cradle, Generate Biomedicines, Xaira, Iambic, Profluent, closed-loop iterative design with experimental feedback.
diffusion models specialized for CDR loops and full antibody scaffolds.
Enzyme design
generative approaches.
AI-guided directed evolution.
the current standard pipeline for designing novel enzymes.
generative approaches.
Nucleic acid design
generative models for aptamer design (still early, the tool you'd want for the sweet-receptor aptamer thought experiment).
variant effect across DNA, RNA, and protein in a single model.
RNA codon optimization for stability and translation efficiency.
generative models for aptamer design (still early, the tool you'd want for the sweet-receptor aptamer thought experiment).
for therapeutic mRNA.
Protein design
alternative protein generative models.
generative approaches.
alternative protein generative models.
alternative protein generative models.
alternative protein generative models.
inverse folding, given a backbone, predict a sequence that folds to it. The standard pairing with RFdiffusion.
denoising diffusion over protein backbones, conditioned on desired binding partners or scaffolds. The dominant tool for de novo binder design.