AI Paper Atlas

Research map / Probabilistic and scientific modeling / Molecules and proteins

EpiCoCo: De Novo Epitope Generation via MHC-Context Co-Modeling and Contrastive Affinity Guidance

Haoyang Luan, Gufeng Yu, Letian Chen, Zhenran Xiao, Yueshan Huang, Junkun Guo, Yang Yang

ICML 2026 · Probabilistic and scientific modeling › Molecules and proteins · approach: Protein fitness and variant effect prediction

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
jointly model MHC and epitope (co-modeling) instead of static MHC conditioning (ablation)EpiCoCo (no ASP, no CAG) · 100 MHCsbinding free energy dG (REU): -32.75 -> -41.62Improved
add Affinity Soft Prompting (continuous affinity conditioning) (ablation)EpiCoCo · 100 MHCsbinding free energy dG (REU): -32.75 -> -41.38 (static); co-modeling -41.62 -> -43.25 or lowerImproved
add Contrastive Affinity Guidance at inference on top of co-modeling + ASP (ablation)EpiCoCo · 100 MHCsbinding free energy dG (REU): -42.67 -> -45.20 (2.53 REU, ~6%)Improved
use EpiCoCo instead of PepFlow (static MHC backbone)PepFlow · 100 MHCsbinding free energy dG (REU): -23.75 -> -45.20Improved
use EpiCoCo instead of RLpMIECRLpMIEC · 100 MHCs vs 13 allelesbinding free energy dG (REU): -37.55 -> -45.20 (23%)Improved
use EpiCoCo with negative-informed generation on alleles with no positive training dataPepPPO, PepPrCLIP · 10 few-negative-shot allelesbinding free energy dG (REU): HLA-B*15:10 -28.06 -> -70.92Improved
use EpiCoCo rather than PepGLAD on diversityEpiCoCo · 100 MHCsdiversity (%): 75.41 (PepGLAD) vs 66.76 (EpiCoCo)Hurt
in silico mutate anchor residue P2 M to A (HLA-E*01:01) (ablation)EpiCoCo generated epitope · HLA-E*01:01ddG (REU): +51.22 REUHurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy