Neural message passing for joint paratope-epitope prediction
read the original abstract
Antibodies are proteins in the immune system which bind to antigens to detect and neutralise them. The binding sites in an antibody-antigen interaction are known as the paratope and epitope, respectively, and the prediction of these regions is key to vaccine and synthetic antibody development. Contrary to prior art, we argue that paratope and epitope predictors require asymmetric treatment, and propose distinct neural message passing architectures that are geared towards the specific aspects of paratope and epitope prediction, respectively. We obtain significant improvements on both tasks, setting the new state-of-the-art and recovering favourable qualitative predictions on antigens of relevance to COVID-19.
This paper has not been read by Pith yet.
Forward citations
Cited by 6 Pith papers
-
EpiFormer: Learning Antigen-Antibody Interactions for Epitope Prediction via Geometric Deep Learning
EpiFormer improves epitope prediction F1 score by over 40% via early-fusion cross-attention in GNN layers and sparsity-aware objectives, while recovering known biology as emergent behavior.
-
ConTact: Contact-First Antibody CDR Design via Explicit Interface Reasoning
ConTact decomposes CDR design into surface fingerprint learning, contact prediction, and contact-gated sequence generation using distance-biased attention and weighted loss, reporting 7% RMSD and 10% F1 gains on CHIME...
-
ConTact: Contact-First Antibody CDR Design via Explicit Interface Reasoning
ConTact introduces a contact-then-act architecture with distance-biased cross-attention and contact-weighted loss for antibody CDR design, reporting 5-6% better backbone RMSD and superior contact metrics on CHIMERA-Be...
-
Deciphering Fingerprints of 3D Molecular Surfaces for Accurate Epitope Prediction
SurfBind applies a Transformer with patch-level surface modeling and binder-aware cross-attention to 3D molecular surfaces, reporting state-of-the-art epitope prediction on SAbDab and DB5.5 with generalization to unse...
-
AgForce Enables Antigen-conditioned Generative Antibody Design
AgForce improves antigen-conditioned antibody design by using framework dropout, gated bottlenecks, hyperbolic cross attention, MDN sequence head with Potts-like coupling, annealed MCL, and antigen cycle consistency t...
-
EvoStruct: Bridging Evolutionary and Structural Priors for Antibody CDR Design via Protein Language Model Adaptation
EvoStruct integrates evolutionary priors from a protein language model with structural priors from an E(3)-equivariant GNN to raise amino acid recovery by 16% and diversity by 2.3x on CHIMERA-Bench while cutting perpl...
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.