Griffith University

Dirichlet flow-matching model for protein design that generates family-aware sequences from ancestral-reconstruction priors, not random noise.

Hierarchical immune cell type annotation for scRNA-seq, rendering expression as images for a CNN. 93.2% mean subtype accuracy over seven datasets.

Full-atom flow matching model that generates a ligand and the induced-fit holo pocket together, starting from an apo binding site.

Protein language model that generates paired heavy and light chain human antibodies from an antigen prompt, with binders validated in vitro.

RNA language model trained on multiple sequence alignments of Rfam families, predicting secondary structure and solvent accessibility from homology.