Design
Designing against fleeting pockets, confirming ligand-dependent protection by HDX-MS, then testing whether that molecular effect produces useful biology.
Designing against fleeting pockets, confirming ligand-dependent protection by HDX-MS, then testing whether that molecular effect produces useful biology.
A generated molecule is a hypothesis. Paired measurement and biological testing decide whether it advances.
Once a transient pocket is resolved, generative models propose candidate molecules against the binding-competent ensemble. Those designs are prioritized for an experimental test that compares the same protein in two matched conditions: protein alone and protein plus molecule.
HDX-MS measures deuterium uptake across both samples. Regions that take up less deuterium in the molecule condition are more protected, showing that the candidate has changed the protein's conformational behaviour. A reproducible protection signature earns the next test: a bespoke assay built around the biology of that program.
The rare states resolved by simulation define a transient pocket that design can act on. Generative models propose candidate molecules against the weighted, binding-competent conformers, then prioritize a focused set of hypotheses for experimental testing.
Each shortlisted molecule is mixed with the target protein. A matched protein-alone control is prepared in parallel, with buffer, protein concentration, and exchange timing held constant so the molecule is the meaningful difference between the two samples.
Both conditions enter the same HDX-MS workflow. Peptide regions with lower deuterium uptake in the protein-plus-molecule sample are more protected, indicating direct or allosteric stabilization. The full pattern, not a single score, determines whether the result is reproducible and worth pursuing.
Candidates with a clear protection signature advance into bespoke biochemical or cellular assays tailored to the program. These assays ask whether the measured change in the protein produces the intended biological effect, and the result informs the next design round.
Generated candidates are prioritized, mixed with the target protein, and compared with a protein-alone control by HDX-MS. Regions with lower deuterium uptake in the molecule condition form a protection signature; confirmed candidates then advance into bespoke biological assays.