KRAS G12D pMHC
KRAS G12D peptide VVVGADGVGK on HLA-A*11:01
Design a mutation-specific pMHC binder
Cells that carry the KRAS G12D mutation can show T cells a ten-residue piece of the mutant protein, VVVGADGVGK, in the groove of the MHC molecule HLA-A*11:01. The wild-type stretch, VVVGAGGVGK, has a glycine where the mutant has an aspartate, so the two peptides differ at one position of ten and sit in the same groove. This challenge asks for three things: bind the G12D complex, do not bind the wild-type one, and do not bind the same HLA with no peptide in it.
KRAS G12D peptide on HLA-A*11:01 · PDB 7OW6
Target information
Design against the whole peptide-MHC complex: the HLA-A*11:01 heavy chain, β2-microglobulin and a ten-residue peptide held in the groove. We recommend targeting the peptide and the groove around it, the surface a T-cell receptor reads. The rest of the complex is the same for both peptides, and the same again with no peptide, so a binder that sits there cannot tell them apart.
KRAS G12D pMHC
To bind- HLA-A*11:01 heavy chain
- Q5S3G3 ↗ 25–305 · 281 aa
- β2-microglobulin
- P61769 ↗ 21–119 · 99 aa
- KRAS G12D peptide
- P01116 ↗ 7–16 · 10 aa
- Assay construct
- Complex of the three chains · tags not shown
- Allele
- A*11:01:01:01 · IPD-IMGT/HLA HLA00043
- Mutation
- KRAS Gly12 to Asp · peptide position 6
- Recommended epitope
- The peptide and the groove around it
- Structure reference
- 7OW6 · Chains A–C · TCR removed ↗
HLA-A*11:01 heavy chain sequence
GSHSMRYFYTSVSRPGRGEPRFIAVGYVDDTQFVRFDSDAASQRMEPRAPWIEQEGPEYWDQETRNVKAQSQTDRVDLGTLRGYYNQSEDGSHTIQIMYGCDVGPDGRFLRGYRQDAYDGKDYIALNEDLRSWTAADMAAQITKRKWEAAHAAEQQRAYLEGRCVEWLRRYLENGKETLQRTDPPKTHMTHHPISDHEATLRCWALGFYPAEITLTWQRDGEDQTQDTELVETRPAGDGTFQKWAAVVVPSGEEQRYTCHVQHEGLPKPLTLRWELSSQPTβ2-microglobulin sequence
IQRTPKIQVYSRHPAENGKSNFLNCYVSGFHPSDIEVDLLKNGERIEKVEHSDLSFSKDWSFYLLYYTEFTPTEKDEYACRVNHVTLSQPKIVKWDRDMKRAS G12D peptide sequence
VVVGADGVGKThree objectives
This challenge consists of 3 main objectives, and each will be accounted for when selecting the best designs.
01
Binding affinity
Design a binder against the KRAS G12D peptide presented by HLA-A*11:01. We recommend targeting the peptide and the groove around it, which is the surface a T-cell receptor reads and the only part that differs from the same HLA carrying another peptide.
We'll measure its binding affinity using the complex of HLA-A*11:01, β2-microglobulin and VVVGADGVGK.
02
Mutation specificity
On top of that, your binder must not bind the same HLA-A*11:01 carrying the wild-type KRAS peptide VVVGAGGVGK. The two peptides are identical except at position 6, residue 12 of KRAS, where the mutant has an aspartate and the wild type a glycine.
A binder that cannot tell them apart cannot single out cells that carry the mutation. For this, we will measure binding against the wild-type complex in the same conditions as the G12D one.
How designs are ranked
Every tested design is measured against all three targets in the same conditions: the KRAS G12D complex, the wild-type KRAS complex and the HLA-A*11:01 with β2-microglobulin that carries no peptide, with the design immobilised and the target in solution. A design counts as binding a target when its sensorgram gives a fittable curve and a KD, or, where the curve cannot be fitted, when the association shift clears 300% over the negative control. Otherwise we call it no detectable binding.
- 1
Mutation specificity
A measurable KD against the KRAS G12D complex and no detectable binding to the wild-type KRAS complex. Designs that bind both rank after these, by the ratio of their KD for the wild-type complex to their KD for the G12D complex, largest first.
- 2
Peptide-dependent binding
A measurable KD against the KRAS G12D complex and no detectable binding to the peptide-free HLA-A*11:01 with β2-microglobulin. Designs that bind it rank after these, by the ratio of their KD for the peptide-free target to their KD for the G12D complex, largest first. A binder that reads the HLA or the β2-microglobulin instead of the peptide binds all three targets and ranks there.
- 3
Affinity
Lowest KD against the KRAS G12D complex. A design that binds the peptide and the groove around it has an advantage. Tell us which surface you target and link your evidence. If you don't, we take it from a predicted complex. A design that binds somewhere else is still measured and ranked.
The winner has to clear all three steps, with ties settled in this order, and each step also awards an honourable mention to the best design that did not win.
Timeline
Oct 26 – Nov 8, 2026
- 1nowSubmit designs
Submissions open
Design, review, and submit. Submissions close Sunday, November 8 at 23:59 Anywhere on Earth (AoE).
- 2
Submissions closed
The window is closed. No further designs are accepted for this challenge.
- 3
Selection in progress
Anthropic and Adaptyv select designs for synthesis using a Claude-based workflow, published after the competition.
- 4
Wet-lab validation in progress
Selected designs go to Adaptyv's automated lab for experimental validation.
Submission requirements
- Length
- 10–250 amino acids
- Format
- Single chain protein, nanobody, scFv, or Fab
- Designs per participant
- Up to 40 for Track 1, up to 20 for Tracks 2 and 3
- Design
- De novo designs only