University of Waterloo

Text-guided protein editing framework with disentangled structure and function latents, edited by rewriting either description at inference time.

DNA barcode foundation model whose masked-autoencoder pretraining keeps mask tokens out of the encoder, for arthropod taxonomic classification.

Antibody sequence design conditioned on antigen structure, generating CDRs or full variable regions without epitope annotation or docked frameworks.

Immune protein structure prediction for TCRs, antibodies, and nanobodies. Adapts ESMFold with LoRA, reaching 1.31 Å RMSD on the CDR3-beta loop.

Histopathology patch encoder giving 1024-dimensional embeddings for slide search and cancer subtyping, trained on TCGA with weak slide labels.