Peptiter / DiscoveryLab
Case studies

Four drug candidates, run through the pipeline — honestly.

We took real, named drugs and asked what the platform can say about them from independent public data. Two are mechanism checks against human disease data; Rabeximod shows how we re-position a safe asset whose Phase 2 missed on duration; Ozempic shows the platform reading a blockbuster two ways — as a certified controller arm and as the foil for an endogenous-GLP-1 alternative. We show the methodology, the result tables, and — deliberately — the weak spots. Nothing here is an efficacy claim.

How to read these

Public-data validations of a mechanistic premise — not efficacy claims.

For each candidate we ask one falsifiable question: does the drug's target behave, in independent public human data, the way its mechanism requires? We pull peer-reviewed datasets, run the analysis transparently, and report what survives and what does not. These are mechanism checks and trial-design critiques — they are not statements that either drug works, and they do not use any sponsor's patient-level data.

What we use
Only public, peer-reviewed sources: SEA-AD brain-cell atlas, GTEx, the CZ CELLxGENE Census, PhysioNet (PADS), PubChem, and published trial papers. Every effect size is sourced; derived or assumed numbers are flagged as such.
What we report
The directions that hold, the strength of the signal, and — explicitly — the caveats a sceptical reviewer would raise first. Modest results are shown as modest. No cherry-picking; out-of-fold numbers only.
What we never claim
That a mechanism check is evidence of clinical efficacy. Target localization, expression direction, and endpoint power are necessary, not sufficient. The clinical questions remain open.
Case study 01 · Alzheimer's / Parkinson's

Buntanetap — do the targets accumulate in the vulnerable neurons?

Buntanetap (ANVS401 / posiphen) is a small molecule proposed to reduce the translation of several aggregation-prone proteins — APP/Aβ, tau (MAPT), α-synuclein (SNCA), and TDP-43 (TARDBP). The mechanistic premise is testable in public human brain-cell data: as disease progresses, do these target transcripts rise in the neurons that actually degenerate? If a target does not accumulate where the damage is, reducing it makes little sense there.

Candidate card
Namesbuntanetap · posiphen · ANVS401 · (+)-phenserine enantiomer
PubChem CID11249342
Formula / MWC20H23N3O2 / 337.4
Sponsor / routeAnnovis Bio / oral (investigational)
Proposed targetsAPP/Aβ · MAPT/tau · SNCA/α-synuclein · TARDBP/TDP-43
Clinical stance12-week Ph2/3 missed its ITT endpoints; signal from a post-hoc pTau217+ mild-AD subgroup. High-signal, high-hype — treated accordingly.
MethodWe staged 84 SEA-AD donors by neuropathology (Braak / Thal / CERAD / LATE) and analysed SEA-AD's published per-cell-type differential-expression along the Continuous Pseudo-progression Score (CPS): for every gene in 139 cortical cell subtypes, the direction of change as Alzheimer's advances. logFC > 0 = target rises with disease = the direction buntanetap's production-reduction would reverse. The neuronal atlas has been public since Feb 2024.
SEA-AD — share of cell subtypes where each target RISES with disease (CPS)
TargetExcitatory neuronsInhibitory neuronsGlia / immune
APP / Aβ76% up31% up (down)50% / 37%
MAPT / tau75% up51% up100% glia (66% sig)
SNCA / α-synuclein89% up83% up71% / 75%
TARDBP / TDP-4373% up46% up33% / 37%
In the excitatory neurons that degenerate in AD, all four targets trend up in a clear majority of ~40 independent subtypes — a consistent direction far beyond chance. SNCA is the most robust, brain-wide disease-up target. APP is cell-type-specific (up in excitatory, down in inhibitory), so the earlier microglia-flat APP was an immune-compartment artifact, not the whole story.
SEA-AD microglia/immune release — per-donor pseudobulk vs aggregation burden (n=84)
Targetcorr. with burdenAD − control
SNCA / α-synuclein+0.36+0.20
TARDBP / TDP-43+0.09+0.01
MAPT / tau+0.09−0.03
APP / Aβ−0.01−0.09
Microglial SNCA rises with severity (the neuroinflammation arm); APP/MAPT/TARDBP are flat in immune cells — expected for neuronal-pathology proteins, and why the neuronal DE above is the decisive test.
Honest limits — what this does NOT show
  • This is mRNA expression versus disease severity — association, not causation, and not proof the drug works. Buntanetap acts on translation, so target-mRNA-up means more substrate, not a measured translation rate.
  • Per-subtype statistical significance is modest; the strength is the consistent sign across many subtypes, not any single test.
  • The clinical programme missed its primary ITT endpoints; the supportive signal is a post-hoc, biomarker-defined subgroup. Cortex (MTG) only; PD validation still needs substantia-nigra data.
Case study 02 · Parkinson's levodopa-induced dyskinesia

Mesdopetam (IRL790) — is the target in the right circuit, and is the trial powered?

Mesdopetam is a preferential dopamine D3-receptor antagonist for levodopa-induced dyskinesia (LID) — a symptomatic, circuit-level drug, not disease modification. There is no disease-progression axis in brain tissue, so the testable premise is target localization: is DRD3 confined to the striatal motor circuit, in the medium spiny neurons that carry the dyskinesia program? We then critique the Phase III design and show how a wearable endpoint changes it.

Candidate card
Namesmesdopetam · IRL-790 · IRL790
PubChem CID70980485
Formula / MWC12H18FNO3S / 275.34
Sponsor / routeIRLAB Therapeutics / oral small molecule
Mechanismpreferential DRD3 antagonism (Ki ≈ 90 nM; ~10× over D2) — D3-preferential, not D3-pure (also D2/σ1/5-HT/NET).
Clinical stancePhase IIb missed its primary endpoint (Good ON-time diary); UDysRS dyskinesia endpoints positive at 7.5 mg BID.
MethodReceptor localization from GTEx v8 bulk RNA-seq (median TPM by brain region) and the CZ CELLxGENE Census (207,260 human striatal cells, cell-type resolved). Phase III power from a noncentral-t / Monte-Carlo simulation anchored on the published Phase IIb effect. All effect sizes sourced; derived SD and reliability assumptions flagged.
Part 1 · Target localization

Is DRD3 in the right circuit?

Two public human datasets — bulk regional (GTEx) and single-cell (CELLxGENE) — locate the target.

GTEx v8 — DRD3 is striatum-restricted (median TPM)
Brain regionDRD3DRD2DRD1
Nucleus accumbens2.9551.931.5
Caudate0.4740.729.5
Putamen0.3645.722.5
Substantia nigra0.008.70.2
Cortex (BA9)0.021.05.7
Cerebellum / hippocampus0.00~1~0.5
DRD3 is 97.7% concentrated in the striatum/basal ganglia and ~0 elsewhere — the most anatomically restricted dopamine receptor (DRD2 spills into substantia nigra, DRD1 into cortex). An anatomically confined target supports a circuit-specific effect ceiling.
CELLxGENE Census — DRD3 is MSN-specific (% of cells positive)
Striatal cell typeDRD1DRD2DRD3
Indirect-pathway MSN (D2)2%99%59%
Direct-pathway MSN (D1)90%17%54%
Inhibitory interneuron5%40%3%
Oligodendrocyte1%1%0.2%
Astrocyte / microglia<1%<1%<1%
DRD3 is positive in ~53–59% of medium spiny neurons but ≤0.8% of glia — a ~70× enrichment into the projection neurons that form the dyskinesia circuit. Honest nuance: DRD3 sits in both D1- and D2-MSNs (broader than the classic "D1-only" model), tempering any direct-pathway-selectivity claim.
Part 2 · Phase III critique → a wearable endpoint

We critique the Phase III design, and show how a wearable endpoint changes it.

The registered Phase IIb primary (a home diary) failed; the design now leans on UDysRS. The left table is the power critique against the real two-study regulatory bar; the right shows how a continuous wearable motor-state readout — the InVivo biomarker — rescues it.

Phase III power — two-study regulatory bar (planned ~65/arm)
Scenario (UDysRS 1+3+4)1-study2-study
Observed Phase IIb Δ=6.20.840.71
+ realistic 20% dropout0.760.58
Winner's-curse −15% (Δ=5.3)0.710.51
Higher Ph3 placebo response0.620.38
If 250–270 were per-study (n=130)0.990.98
At the observed effect the two-study bar is only 0.71 (≈0.58 with dropout). Δ=6.2 is the selected survivor of a failed-primary trial; a 15% shrink makes the two-study win a coin-flip. Cheapest fix: per-study n ≈ 80–100/arm.
Digital endpoint — a wearable readout rescues the power
Endpoint (reliability R)2-studyn/arm 80%
Home diary — the FAILED primary (0.40)0.46113
UDysRS clinic scale (0.60)0.7176
Wearable, 1-day capture (0.80)0.8657
Wearable, full-period (0.90)0.9151
A more reliable endpoint de-attenuates the effect (d_obs = d_true·√R). The diary's low reliability quantitatively explains why the Phase IIb primary failed; a continuous wearable index lifts two-study power 0.71 → 0.91, or ~33% fewer patients. Reliability values are flagged assumptions.
Motor-state model — validated on real Apple Watch data (PADS)
Model (subject-grouped CV)AUCsensspec
Logistic (deployed, native Swift)0.760.670.71
Random forest (server option)0.840.680.81
156 subjects (78 PD / 78 healthy), 100 Hz wrist accel+gyro, 28 motor features (incl. tremor/dyskinesia bands, sample-entropy, Poincaré). PD-vs-healthy — NOT a dyskinesia classifier (that needs gated labels). Swift features match the Python reference to <1%. The front-end is real; the dyskinesia head is future work.
Honest limits — what this does NOT show
  • Localization is necessary, not sufficient, and the molecule is D3-preferential, not D3-pure — the antidyskinetic effect cannot be attributed to D3 alone.
  • GTEx/Census show the healthy baseline; the LID-driving dorsal D3 pool is partly induced by L-DOPA (Bordet 1997) and is not visible here.
  • PADS labels are diagnosis, not dyskinesia state; the power gain is a design argument contingent on the wearable index actually tracking the clinical construct. The Phase IIb primary miss stands.
Part 3 · Developability & competitive position

Is it drug-like, and how does it stack up?

CNS chemistry and the off-target window, then the standardized effect against the approved benchmark — told straight.

ADMET — CNS developability & the off-target window
Metricmesdopetamamantadine
CNS MPO (Wager; ≥4 good)4.834.13
DRD3 ligand efficiency0.55
D3 → D2 selectivity~9×
D3 → σ1 selectivity~7×
hERG / CYP / P-gpnot public
Chemistry is drug-like and BBB-plausible (CNS MPO 4.8, zero Lipinski violations). The honest caution is the narrow ~7–9× window over D2/σ1 — at therapeutic D3 occupancy there is real D2/σ1 engagement, so the effect is not cleanly D3-mediated. hERG/CYP/P-gp are not publicly measured (stated, not guessed).
Competitive — standardized vs the approved benchmark (amantadine ER)
Therapy / endpointCohen's d% scale
mesdopetam — UDysRS 1+3+40.537.0%
amantadine ER — EASE LID0.647.6%
amantadine ER — EASE LID 31.1613.8%
On standardized effect, mesdopetam is at or below amantadine ER — and amantadine's was a succeeding primary endpoint, while mesdopetam's was a secondary after the primary failed. The differentiation is tolerability (Phase IIb AE profile broadly placebo-like), not efficacy — and that edge is unproven against an active comparator. Not a head-to-head trial; sub-scales differ.
Part 4 · Disease-state DRD3

Does the often-cited mechanism actually hold?

A deliberate stress-test of the convenient story — reported even though it does not confirm it.

Disease-state DRD3 — is the target INDUCED in dyskinesia? (a non-confirmatory check)
Gene (dyskinetic vs control)rat striatummouse D1-MSNverdict
DRD3 (the target)+0.43−0.79modest in rat, down in mouse
Fosb / ΔFosB (LID control)+1.96+3.93✓ program captured
Pdyn (LID control)+1.66+1.71✓ program captured
Arc (LID control)+1.23+3.40✓ program captured
log2 fold-change in 6-OHDA+L-DOPA dyskinesia (GSE88726 rat bulk; GSE55096 mouse cell-type bacTRAP). The often-cited "ectopic DRD3 induction drives LID" story (Bordet 1997) is only modestly and inconsistently borne out — a small dyskinesia-specific rise in rat striatum (~1.35×) but absent/reversed in the mouse D1-MSN translatome, even though the LID program (ΔFosB/Pdyn/Arc) is textbook- perfect in both. This corrects an over-cited mechanism rather than undermining the drug: mesdopetam's basis is the constitutive DRD3 it antagonizes (shown above), not a disease-induced pool.
Case study 03 · Rheumatoid arthritis → fibrosis

Rabeximod — a safe macrophage-axis small molecule that missed on duration, not mechanism.

Rabeximod is a real, de-risked oral small molecule: well-tolerated in 200+ patients, with a mechanism that impairs differentiation and activation of pro-inflammatory macrophages and dendritic cells downstream of TLR2/TLR4. Its Phase 2 in rheumatoid arthritis missed the primary endpoint — but the benefit was emerging at 16 weeks, suggesting the trial was simply too short. That is the exact signature DiscoveryLab is built to read: a mechanistically-sound monotherapy that under-controls the disease variable within the trial's window. We don't own this asset — but rather than stop at re-reading the study, we ran it through the pipeline: routed it through ChemCheck, added it as a fourth orthogonal arm to our certified IL-6 controller, and minted a real CI-checked receipt (lyapunov_rabeximod_ra_combination). The computed result, below, is the part no press release contains.

Candidate card
Namesrabeximod · CYX-001 · ROX9 · DB05772
Class / routeoral small molecule (investigational)
SponsorsOxyPharma → Cyxone → Renity Pharma
Mechanismimpairs monocyte→DC / pro-inflammatory macrophage differentiation; suppresses activation downstream of TLR2/TLR4. Renity now frames it as a first-in-class PGK1 inhibitor (immuno-metabolic) — a mechanism ambiguity worth resolving.
Indicationsrheumatoid arthritis → pivot to chronic kidney disease / fibrosis
Clinical stancePhase 2 RA (n≈224) missed the primary endpoint at 12 weeks; benefit emerging at 16 weeks. Safe in 200+ patients. A Phase 2b at 24 weeks was the proposed fix.
MethodWe route Rabeximod through the small-molecule developability lane (ChemCheck) and place it as the macrophage-differentiation mechanism point on the rheumatoid-arthritis inflammatory axis — orthogonal to the cytokine arms in our already-certified IL-6 controller (lyapunov_il6_combination). Then we ask the controller question: does a macrophage arm ⊕ a cytokine arm reach control adequacy where each single agent under-controls — the same logic that the 16-week signal hints at.
Public clinical record — the 'miss' is a duration / under-control signal
ReadoutValueController reading
Phase 2 RA, patientsn ≈ 224powered cohort
Primary endpoint @ 12 wknot metunder-controls in-window
Effect @ 16 wkemerging benefitcontrol adequacy reached late
Safety, exposed patients200+, well-toleratedwide therapeutic window
Proposed fixPhase 2b @ 24 wkextend the dosing cycle
A safe agent whose benefit appears only past the trial window is the macroscopic version of kMono < kMin — the monotherapy under-controls across the real cycle. Our framework names that failure in advance and asks whether an orthogonal arm closes the gap.
Where DiscoveryLab places it — mechanism points on the RA inflammatory axis
ArmModalityMechanism pointStatus
IL-6 trap (sgp130-Fc)designed proteinligand reductioncertified
JAK inhibitorsmall moleculedownstream signalingcertified
Designed receptor-block peptidepeptidereceptor blockcertified
Rabeximodsmall moleculemacrophage / cytokine sourcecertified
The first three are the shipped lyapunov_il6_combination receipt. Rabeximod adds a distinct, orthogonal mechanism point (upstream cellular source), and the four-arm controller is now its own CI-checked receipt: lyapunov_rabeximod_ra_combination.
What the platform COMPUTED — not in the study (4-arm vs the IL-6 trio)
Computed outputIL-6 trio+ Rabeximod
Closed-loop K (peak / trough)0.131 / 0.1090.141 / 0.124
Monotherapy (each arm alone)K < 0.100K < 0.100 (under-controls)
Robustness (81-world ±15%)79%89%
Tolerated uniform potency drop−10%−28%
Dominant edge failureunder-control @ troughover-suppression @ peak
Pre-registered factorialconfirm / null→killconfirm / null→kill
This is the answer to "what do you add beyond reading the study." The four-arm controller is machine-checked stable, more robust (89% vs 79%), and tolerates far more surrogate error (−28% vs −10%) — but the failure mode inverts: adding a fourth immunosuppressive arm trades the trough under-control risk for over-suppression at peak, so the design lever becomes de-escalation, not extension. None of this is in any press release — peptiter-reproduce-combination regenerates it.
For the developer of Rabeximod

Three things this gives you that the trial never could.

Your Phase 2 established that Rabeximod is safe and that an effect was emerging at 16 weeks. The questions it left open — does that late signal become adequate control in combination, how fragile is that to the numbers, and what is the right regimen — are exactly the ones DiscoveryLab answers as computed, auditable artifacts. Everything below regenerates from one command and recompiles in CI; none of it is a black box.

01

A computed verdict, not a narrative.

The platform turns "benefit emerging at 16 weeks" into a predicate the proof engine either discharges or rejects. Rabeximod on its own computes to K < 0.100 — it under-controls, the formal shadow of the missed 12-week endpoint. Added as a fourth orthogonal arm to a certified IL-6 trio, the combined controller lands K = 0.141 at peak and holds 0.124 at the dosing trough, inside the 0.10–0.16 control band, with a machine-checked Lean proof that the closed loop is stable across the entire cycle (lyapunov_rabeximod_ra_combination). That is a yes/no you can carry into a partnering or trial-design conversation — not a hope that a longer trial works out.

02

A quantified robustness margin — on your own surrogate numbers.

The fair objection to any model is "the efficacy and PK inputs are guesses." We answer it numerically. Across an 81-world grid that perturbs every agent's potency and trough retention by ±15%, the Rabeximod-armed controller stays valid in 89% of worlds — up from 79% for the cytokine trio alone — and it absorbs a uniform −28% potency drop before it breaks, versus −10% without it. In plain terms, adding Rabeximod's orthogonal mechanism makes the regimen markedly less fragile to being wrong about the numbers. That robustness delta is the quantified case for the asset's place in the combination — not an assertion that it helps.

03

A falsifiable design instruction — including where it would fail.

The most useful output is the one no press release contains: where the combination breaks, and what to do about it. Adding Rabeximod inverts the dominant failure mode — the cytokine trio's weak point is under-control at the trough; with Rabeximod's fourth suppressive arm it flips to over-suppression at peak (an immunosuppression / infection ceiling, not loss of effect). The design instruction follows directly: a Rabeximod regimen should de-escalate the cytokine arms (lower dose / longer interval), not simply extend duration the way the monotherapy's 16-week signal implied. And it ships with a pre-registered factorial whose kill-criterion is named in advance — modeled arm → CONFIRM, the no-super-additivity null → KILL — so the next experiment is decision-grade, not exploratory.

Shareable section · /case-studies#rabeximod-value. Every number regenerates from peptiter-reproduce-combination and recompiles under lake build in CI — auditable by your team, line by line. The Rabeximod inputs above are documented surrogates; send us your measured efficacy and PK and the whole certificate re-mints from one command (remint-rabeximod.sh) on your numbers.
Honest limits — what this does NOT show
  • Rabeximod is not our asset. Everything above is a re-analysis of the public record (OxyPharma / Cyxone / Renity); we make no efficacy or safety claim of our own.
  • The minted receipt proves the math is consistent given the model and surrogate efficacy/PK numbers — it is an in-silico certificate, not evidence the combination works in patients. The wet-lab factorial is still the arbiter.
  • The mechanism is ambiguous in public sources — macrophage/dendritic-cell differentiation (TLR2/4) vs the newer PGK1 / immuno-metabolic framing. Which one attributes to the RA endpoint is a signed-attribution question, not a settled fact.
  • The RA→CKD/fibrosis pivot is plausible (macrophage-driven inflammation feeds fibrosis) but is a separate axis that would need its own controlled-variable and endpoint before any claim.
Case study 04 · Obesity / Type-2 diabetes

Ozempic (semaglutide) — the certified controller arm, and the endogenous-GLP-1 alternative.

Ozempic is the most consequential metabolic drug of the decade — and it is already woven through DiscoveryLab two ways, both from public data. It is the GLP-1 arm of our shipped, machine-checked glucose controller; and it is the foil for a rank-only program that asks whether the gut can be steered to raise the body's own meal-triggered GLP-1 instead of injecting an agonist. We don't own this asset and we dispute none of its efficacy; we show what the platform can say about it that a label can't.

Candidate card
Namessemaglutide · Ozempic · Wegovy · Rybelsus
Class / routeGLP-1 receptor agonist peptide; SC weekly / oral daily
SponsorNovo Nordisk
Mechanismlong-acting GLP-1R agonist — glucose-dependent insulin secretion, slowed gastric emptying, central satiety. Continuous, supraphysiologic, exogenous receptor agonism.
Indicationstype-2 diabetes · chronic weight management
Clinical stanceapproved, highly effective, blockbuster. Real-world frictions: GI tolerability, lean-mass loss, and weight regain on discontinuation.
MethodTwo honest readings, both from public data. Reading 1: semaglutide is the GLP-1 arm of our certified glucose controller (lyapunov_glucose_combination) — orthogonal to an insulin analog and an SGLT2 inhibitor, with a machine-checked hypoglycemia ceiling. Reading 2: a rank-only, measurement-first program models steering the gut ecosystem to raise endogenous, meal-triggered GLP-1 — a closed-loop alternative to flooding the receptor.
Reading 1 — Ozempic as a certified, hypoglycemia-bounded controller arm
ArmModalityMechanism pointStatus
Semaglutide (Ozempic)peptideincretin secretioncertified arm
Insulin analog (basal)designed proteindirect effectorcertified
SGLT2 inhibitorsmall moleculerenal clearancecertified
The shipped, CI-recompiled lyapunov_glucose_combination receipt: closed-loop gain K = 0.126 / 0.108 (peak / trough) in-band, robust in 74% of an 81-world ±15% grid, with the over-suppression ceiling = hypoglycemia, proven. The point is composition: three orthogonal arms reach control at a tolerated dose of each — the formal version of dose-sparing, which is exactly the lever against Ozempic's GI burden.
Reading 2 — the endogenous-GLP-1 alternative (rank-only, measurement-first)
AxisHost receptorLever (diet / substrate)Rank*
PropionateFFAR2/3fermentable fiber (arabinoxylan / β-glucan)0.81
Secondary bile acidsTGR5bile-acid guild (map before steering)0.80
ButyrateFFAR2/3 + barrierresistant starch + inulin0.83
Raise the body's own meal-triggered GLP-1 by steering cross-feeding guilds, rather than injecting an agonist — a closed-loop, physiologic control philosophy. *Rank scores are rank-only ordering, not probabilities of clinical effect. Diet/substrate steering only — no live or engineered organism — and explicitly not an Ozempic substitute: this raises endogenous GLP-1 modestly, not to pharmacologic levels.
The controller reading

Ozempic is high-gain, open-loop, exogenous receptor agonism — supraphysiologic and continuous (effective, but the source of GI burden, lean-mass loss, and regain-on-stop). The platform contributes the two things a label can't: a proven hypoglycemia-bounded combination that reaches control at tolerated doses (dose-sparing), and a falsifiable, pre-registered measurement plan for the endogenous, meal-responsive alternative. Both are in-silico / rank-only — the bench is still the arbiter.

Honest limits — what this does NOT show
  • Ozempic is not our asset (Novo Nordisk). This is a public-record analysis; we make no efficacy or safety claim, and dispute none of semaglutide's — it is highly effective.
  • The certified glucose receipt proves the math is consistent given documented surrogate efficacy / PK numbers — an in-silico certificate, not evidence about Ozempic in patients.
  • The endogenous-GLP-1 program is rank-only and measurement-first: a falsifiable hypothesis + readouts, explicitly not an Ozempic substitute and not a clinical claim. Scores order axes; they are not effect sizes.
Data provenance & reproducibility

Every number traces to a public source.

The analyses are scripted and reproducible; raw atlases are never redistributed, only the small derived results.

Sources
DatasetUse
SEA-AD (brain-map.org)Buntanetap — neuronal & microglial differential expression along AD pseudo-progression
GTEx v8 (gtexportal.org)Mesdopetam — DRD3 regional localization across the brain
CZ CELLxGENE Census 2025-01-30Mesdopetam — DRD3 by striatal cell type (207k human cells)
PADS (PhysioNet, Varghese 2024)Mesdopetam — Apple Watch motor-state model (real wrist sensor data)
Antonini 2025, Mov Disord Clin PractMesdopetam — Phase IIb effect sizes for the power simulation
Pahwa 2017 / Oertel 2017 (EASE LID)Mesdopetam — amantadine ER benchmark for the competitive comparison
Waters/Sonesson 2020 JPET; Wager 2010Mesdopetam — off-target binding (Ki) and the CNS MPO algorithm
GSE55096 (Heiman) / GSE88726 (Smith)Mesdopetam — disease-state DRD3 in 6-OHDA+L-DOPA dyskinesia (GEO)
PubChemBoth — CNS-developability chemistry
Methods and scripts live in the DiscoveryLab experiments tree (experiments/buntanetap, experiments/mesdopetam). The mesdopetam motor-state model is implemented in the InVivo app (MotorState/).
DisclaimerThese case studies are mechanism and trial-design analyses on public data. They are not clinical claims, not investment advice, and not affiliated with or endorsed by the named sponsors. No patient-level trial data was used. Each result is reported with its caveats so the limitations are visible alongside the findings.