Loading…
Linear B-cell epitope prediction using phylogeny-aware fine-tuning of ESM embedders, with taxon-specific models for 19 pathogen groups.
TCR-peptide binding prediction that fuses ESM-1b receptor embeddings with MolFormer SMILES embeddings through multi-head cross-attention.
Protein binding affinity prediction from sequence alone, returning pKd and per-residue interface labels instead of a yes-or-no interaction call.
Per-residue pKa prediction from sequence alone, a thin MLP head on frozen ESM-2 embeddings reaching 0.48 RMSE across six titratable residue types.
Per-residue pKa prediction from sequence alone, a thin MLP head on frozen ESM Cambrian embeddings that holds up better on engineered buried variants.
Topology-guided protein backbone generation that turns hand-drawn 3D curves into designable structures by steering a diffusion sampler with a sketch.
Whole-genome bacterial pathogenicity prediction from ProtT5 embeddings, alignment-free and taxonomy-agnostic, with per-protein attention scores.
Pocket-based molecular docking with an SE(3)-equivariant diffusion transformer. Places 77.65% of top-1 poses within 2 Å RMSD on PoseBusters.
Multitask binding site prediction across protein, DNA/RNA, ligand, lipid, and ion partners, combining protein language models with equivariant GNNs.
Protein-protein interface embedding model built on Delaunay graphs, reused frozen for antibody-antigen affinity and antibody viscosity prediction.
All-atom generative model for protein complexes that designs multi-chain binders from scratch and performs multimer folding and inverse folding.
E3 ubiquitin ligase-substrate interaction prediction from a LoRA-adapted protein language model fused with structure and subcellular localization.
Multimodal diffusion protein language model co-generating sequence and structure. Bit-level structure supervision cuts folding RMSD from 5.52 to 2.36.
Protein sequence design by flow matching in a compressed language-model latent space, spanning peptides, antibodies, and antimicrobial peptides.
Peptide language model pretrained from scratch on short UniProt sequences, matching ESM-2 on 8 of 9 bioactive peptide tasks at 4.9M parameters.
Spider silk protein language model that generates MaSp repeat sequences from target mechanical properties and predicts those properties from sequence.
Bacterial exotoxin classifier over frozen ProtT5 embeddings that separates secreted toxins from non-toxic secreted proteins at MCC 0.94.
Protein language model interpretability adapter that factors ESM2 and ProtBERT embeddings into named biochemical features plus a residual subspace.
Flexible protein-ligand docking and binding affinity prediction from an apo receptor structure and ligand SMILES, using an 8-layer pair transformer.
Machine-learned interatomic potential supplying quantum-quality geometric restraints for refining cryo-EM and X-ray protein structures in Phenix.
TCR-pMHC specificity prediction that folds frozen ESM-2 and AlphaFold2 representations into a three-body peptide-MHC-CDR3 energy tensor.
Protein binder design that inverts the frozen Boltz-1 all-atom predictor, targeting small molecules, nucleic acids, metals, and modified residues.