An integrated in silico-in vitro workflow for discovering high-affinity, selective antibodies to the KRAS(G12D)-MHC I complex
Ahn, S.; Oh, T.-S.; Suh, S.; Jeon, J.-y.; Lah, S.; Ryu, K.; Kim, H.; Lee, E. G.; Lee, H.; Lee, J.; Kim, D.-K.; Ku, B. M.; Jung, W.; Ahn, M.-J.; Jung, J. U.; Kim, Y.-S.; Oh, B.-H.; Jeong, B.-S.
Show abstract
Antibodies that recognize peptide-loaded class I major histocompatibility complex (pMHC I) molecules could enable therapeutic targeting of intracellular oncogenic proteins, yet their discovery has been hampered by the small size of peptide antigens and allele-specificity. We describe an integrated in silico-in vitro workflow for generating high-affinity, selective antibodies to KRAS(G12D)10 presented by HLA-C*08:02, a clinically validated cancer neoantigen. In silico, multiple human antibody-derived variable fragments (Fvs) plausibly docked to the target pMHC were generated, followed by limited complementarity-determining region (CDR) sequence design. In vitro, CDR diversity was introduced at 3-4 positions per Fv to construct yeast surface display library for iterative selections. This workflow yielded antibodies with exclusive binding to KRAS(G12D)10/HLA-C*08:02 without cross-reactivity. Affinity maturation achieved nanomolar dissociation constants, and incorporation into chimeric antigen receptor T cells enabled specific activation against target-positive cells. This study establishes a practical design-to-function pipeline for TCR-like antibody discovery, and demonstrates the feasibility of therapeutic targeting against KRAS(G12D)-driven malignancies.
Matching journals
The top 6 journals account for 50% of the predicted probability mass.
Similar papers in this journal
Similar papers in this journal
Similar papers in this journal
- Phage display enables machine learning discovery of cancer antigen specific TCRs 96%
- Engineered ACE2-Fc counters murine lethal SARS-CoV-2 infection through direct neutralization and Fc-effector activities 95%
- Elucidating the Characteristics and Clonal Evolutionary Trajectory of Influenza Neuraminidase Broadly Reactive B Cell 94%
Similar papers in this journal
- Beyond Sequence Similarity: ML-Powered Identification of pHLA Off-Targets for TCR-Mimic Antibodies Using High Throughput Binding Kinetics 96%
- Potent SARS-CoV-2 binding and neutralization through maturation of iconic SARS-CoV-1 antibodies 94%
- In silico proof of principle of machine learning-based antibody design at unconstrained scale 94%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.