Pharma exemplar · ADMET Lead Desk · Stage C2 CLOSED PASS · IBM Heron r2 · ibm_marrakesh

Seventeen ADMET logics. One quantum architecture. WOW DONE.

Stage A→C closed on IBM Heron r2 · ibm_marrakesh: same-day full pass. Raw junction gap 0.185 · Aer retention 0.749 · separable contrast 0.184. Locked junction_ryy · 156 qubits · jobs d9qubi7pemts73crsa2g+. Classical methods on the same frozen table already disagree. QPC keeps those logics co-resident and returns a measured architecture scorecard with public IBM job IDs. Public WOW article →

What this job is

In early drug discovery a molecule is judged by many ADMET properties at once: a multi-logic landscape. Those properties pull in different directions. Classical practice weights them into one number, or applies hard filters one after another. This job puts those logics together inside one auditable quantum run on real hardware.

Goal

Freeze a public-proxy desirability table (9 compounds × 17 TDC-aligned ADMET endpoints), encode it as a polycontextural landscape (9 × 17 data qubits + absorb / clear / safety junctions = 156 qubits), run coupled vs no_cross control under locked junction_ryy geometry on IBM Heron r2 · ibm_marrakesh, and read a provisional lead ranking from the coupled arm, compared with classical baselines on the same numbers.

Relevance

Pharma teams already live with multi-objective ADMET conflict (permeability vs toxicity, CNS exposure vs CYP liability, oral absorption vs DILI risk). This exemplar is a bounded, public instance of multi-logic decisioning with IBM job IDs: the shape of an Exemplar Replay a partner can later fill with their own scores under the same schema.

What this is

A bounded, public multi-logic ADMET architecture run: frozen public-proxy desirabilities, locked junction_ryy geometry, coupled vs control vs separable arms, raw junction scorecard, and IBM job IDs. A partner can fill the same schema with their own scores. See the database page for the frozen table and provenance.

How the computation is done

1
Frozen ADMET table
2
17 contextures x 9
3
Heron r2 · 156Q
4
Coupled vs control
5
Lead readout
Steps
Stage C2Heron r2 · ibm_marrakesh
C2 jobsd9qubi7pemts73crsa2g+
Shots8192 / variant
Geometryjunction_ryy · 156Q
Primary metricraw_junction_gap
WOW statusCLOSED PASS

Results (understandable)

(1) Classical methods disagree on this table: that is why multi-logic architecture matters. (2) Stage C2 on IBM Heron r2 · ibm_marrakesh cleared every pre-registered bar the same day. That is the WOW closure.

WOW
Stage A→C CLOSED
0.185
Raw junction gap · Marrakesh
0.749
Aer retention · Marrakesh
0.184
Separable contrast
Hardware compound ranking (Marrakesh C2 coupled arm, provisional lead score)
Rank Compound Class Lead score
1caffeinecns_tool0.736
2acetaminophenanalgesic0.669
3propranololcardiovascular0.588
4ibuprofennsaid0.572
5fluoxetinecns0.568
6terfenadinewithdrawn0.567
7digoxincardiovascular0.530
8cisplatinoncology0.432
9doxorubicinoncology0.343

Reading this: favorable oral / CNS-friendly profiles rise; high-toxicity oncology agents fall. A consistent encoding readout of the multi-ADMET landscape on the frozen public-proxy table.

Classical baselines on the same frozen table

Equal-weight sum ranks: caffeine, acetaminophen, propranolol, ibuprofen, fluoxetine, digoxin, terfenadine, cisplatin, doxorubicin.

Sequential hard filters (hERG, Ames, DILI, P-gp, bioavailability gates) keep: caffeine, propranolol, ibuprofen, fluoxetine. Acetaminophen (weighted-sum #2) drops at the DILI gate. Same table, different classical verdict.

Pareto front (maximize all 17): caffeine, fluoxetine, acetaminophen. Kendall tau (weighted-sum vs sequential) = 0.50: measured disagreement among classical methods.

Hardware ranking order differs from classical equal-weight in places. That difference is an architecture signal on the same table, read together with the raw junction scorecard.

WOW closure (Stage C2 · IBM Heron r2 · ibm_marrakesh): Same-day full pass under locked junction_ryy + no_cross: gap 0.185 · retention 0.749 · separable contrast 0.184. Jobs d9qubi7pemts73crsa2g+ / separable d9qubk8pdb6s73e45q3g. Primary metrics on raw counts. IBM Heron r2 · ibm_fez Stage B (raw gap ~0.18) is supporting evidence on the same desk family. Public article: QPC_ADMET_WOW_ARTICLE.html.

What this demonstrates

Joint multi-logic ADMET decisioning under a published conflict table; co-resident logics in one coherent quantum job; raw junction architecture metrics; Exemplar Replay schema; IBM audit trail.

Frozen score table and column definitions: ADMET_LEAD_DATABASE.html