A ranked CSV is not the finish line. Discovery Launch is built around a harder benchmark: send Om a protein sequence, let LULA-2 rank roughly 40 million molecules, physically obtain the top 6,000, and measure what actually bound.
The customer input is intentionally small: a protein sequence. No structure, no assay, no prior binder set. One sequence starts the loop.
How the challenge runs
First, Om uses LULA-2 to score roughly 40 million molecules against the submitted target sequence. The model is designed for rapid protein-conditioned ranking, so the useful question is not whether it can run one expensive prediction. The useful question is whether it can move likely binders into the part of the list a team can actually test.
Second, the head of the ranking becomes real. Om obtains the top 6,000 molecules for the program and advances them into physical testing.
Third, Om runs affinity-selection mass spectrometry across the validation set. The result is not just another score table. It is a measured readout showing which molecules bound under the assay conditions.
A score is a hypothesis. A binder is a result.
Anyone can hand you a ranked file. The hard part is turning the top of that ranking into compounds, running the assay, and showing what came back.
That is why Discovery Launch is a challenge rather than a model demo. The proof is the closed loop: sequence in, LULA-2 ranking, top-ranked molecules obtained, AS-MS measurement, validated binders returned.
What you get back
When results are ready, the Discovery Launch order page will show the ranked output artifact for the submitted protein. The result package is intended to include structures, LULA scores, ranks, and the measurement behind each validated molecule.
That makes the program useful for two jobs at once: finding immediate binders for a target, and creating a clean validation readout for teams that want to evaluate LULA on their own proteins.
The academic path
Academic validation participants place the Discovery Launch order with Wallet Credits, then reach out to Om for credit-back approval and protein shipment instructions. After Om receives and accepts 200 ug purified protein for the placed order, Om credits back 500,000 Wallet Credits to the account.
This keeps the program simple: one protein sequence starts the computational run, and the purified protein sample helps Om complete the physical validation loop.
Start Discovery Launch, or contact Om for protein credit-back instructions.
